topjohnwu/ndk-busybox
1
0
Fork 0
mirror of https://github.com/topjohnwu/ndk-busybox.git synced 2024-11-24 12:19:49 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
3b84dfc2a8
ndk-busybox/shell/msh.c

5489 lines
102 KiB
C
Raw Normal View History

This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* vi: set sw=4 ts=4: */
/*
* Minix shell port for busybox
*
* This version of the Minix shell was adapted for use in busybox
Update a bunch of docs. Run a script to update my email addr.
2003-07-14 21:21:08 +00:00
* by Erik Andersen <andersen@codepoet.org>
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
*
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
* - backtick expansion did not work properly
* Jonas Holmberg <jonas.holmberg@axis.com>
* Robert Schwebel <r.schwebel@pengutronix.de>
Update a bunch of docs. Run a script to update my email addr.
2003-07-14 21:21:08 +00:00
* Erik Andersen <andersen@codepoet.org>
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
*
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
*
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
* Original copyright notice is retained at the end of this file.
*/
#include <ctype.h>
#include <dirent.h>
#include <errno.h>
#include <fcntl.h>
#include <limits.h>
#include <setjmp.h>
#include <signal.h>
#include <stddef.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <time.h>
#include <unistd.h>
#include <sys/stat.h>
#include <sys/times.h>
#include <sys/types.h>
#include <sys/wait.h>
#include "cmdedit.h"
#include "busybox.h"
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* Conditional use of "register" keyword */
#define REGISTER register
/*#define MSHDEBUG 1*/
#ifdef MSHDEBUG
int mshdbg = 0;
#define DBGPRINTF(x) if(mshdbg>0)printf x
#define DBGPRINTF0(x) if(mshdbg>0)printf x
#define DBGPRINTF1(x) if(mshdbg>1)printf x
#define DBGPRINTF2(x) if(mshdbg>2)printf x
#define DBGPRINTF3(x) if(mshdbg>3)printf x
#define DBGPRINTF4(x) if(mshdbg>4)printf x
#define DBGPRINTF5(x) if(mshdbg>5)printf x
#define DBGPRINTF6(x) if(mshdbg>6)printf x
#define DBGPRINTF7(x) if(mshdbg>7)printf x
#define DBGPRINTF8(x) if(mshdbg>8)printf x
#define DBGPRINTF9(x) if(mshdbg>9)printf x
int mshdbg_rc = 0;
#define RCPRINTF(x) if(mshdbg_rc)printf x
#else
#define DBGPRINTF(x)
#define DBGPRINTF0(x)
#define DBGPRINTF1(x)
#define DBGPRINTF2(x)
#define DBGPRINTF3(x)
#define DBGPRINTF4(x)
#define DBGPRINTF5(x)
#define DBGPRINTF6(x)
#define DBGPRINTF7(x)
#define DBGPRINTF8(x)
#define DBGPRINTF9(x)
#define RCPRINTF(x)
#endif /* MSHDEBUG */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* -------- sh.h -------- */
/*
* shell
*/
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define LINELIM 2100
#define NPUSH 8 /* limit to input nesting */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Fixup all the horrible broken function prototypes -Erik
2002-12-11 07:42:46 +00:00
#undef NOFILE
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define NOFILE 20 /* Number of open files */
#define NUFILE 10 /* Number of user-accessible files */
#define FDBASE 10 /* First file usable by Shell */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* values returned by wait
*/
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define WAITSIG(s) ((s)&0177)
#define WAITVAL(s) (((s)>>8)&0377)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define WAITCORE(s) (((s)&0200)!=0)
/*
Larry Doolittle writes: This is a bulk spelling fix patch against busybox-1.00-pre10. If anyone gets a corrupted copy (and cares), let me know and I will make alternate arrangements. Erik - please apply. Authors - please check that I didn't corrupt any meaning. Package importers - see if any of these changes should be passed to the upstream authors. I glossed over lots of sloppy capitalizations, missing apostrophes, mixed American/British spellings, and German-style compound words. What is "pretect redefined for test" in cmdedit.c? Good luck on the 1.00 release! - Larry
2004-04-14 17:51:38 +00:00
* library and system definitions
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
typedef void xint; /* base type of jmp_buf, for not broken compilers */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* shell components
*/
#define QUOTE 0200
#define NOBLOCK ((struct op *)NULL)
#define NOWORD ((char *)NULL)
#define NOWORDS ((char **)NULL)
#define NOPIPE ((int *)NULL)
/*
* Description of a command or an operation on commands.
* Might eventually use a union.
*/
struct op {
Run msh through indent
2004-08-04 19:16:54 +00:00
int type; /* operation type, see below */
char **words; /* arguments to a command */
struct ioword **ioact; /* IO actions (eg, < > >>) */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op *left;
struct op *right;
Run msh through indent
2004-08-04 19:16:54 +00:00
char *str; /* identifier for case and for */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
Run msh through indent
2004-08-04 19:16:54 +00:00
#define TCOM 1 /* command */
#define TPAREN 2 /* (c-list) */
#define TPIPE 3 /* a | b */
#define TLIST 4 /* a [&;] b */
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define TOR 5 /* || */
Run msh through indent
2004-08-04 19:16:54 +00:00
#define TAND 6 /* && */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define TFOR 7
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define TDO 8
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define TCASE 9
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define TIF 10
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define TWHILE 11
#define TUNTIL 12
#define TELIF 13
Run msh through indent
2004-08-04 19:16:54 +00:00
#define TPAT 14 /* pattern in case */
#define TBRACE 15 /* {c-list} */
#define TASYNC 16 /* c & */
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* Added to support "." file expansion */
#define TDOT 17
/* Strings for names to make debug easier */
Reduce exported from msh applet
2005-09-22 14:38:17 +00:00
#ifdef MSHDEBUG
static char *T_CMD_NAMES[] = {
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
"PLACEHOLDER",
"TCOM",
"TPAREN",
"TPIPE",
"TLIST",
"TOR",
"TAND",
"TFOR",
"TDO",
"TCASE",
"TIF",
"TWHILE",
"TUNTIL",
"TELIF",
"TPAT",
"TBRACE",
"TASYNC",
"TDOT",
};
Reduce exported from msh applet
2005-09-22 14:38:17 +00:00
#endif
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* actions determining the environment of a process
*/
#define BIT(i) (1<<(i))
Run msh through indent
2004-08-04 19:16:54 +00:00
#define FEXEC BIT(0) /* execute without forking */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#if 0 /* Original value */
#define AREASIZE (65000)
#else
#define AREASIZE (90000)
#endif
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* flags to control evaluation of words
*/
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define DOSUB 1 /* interpret $, `, and quotes */
#define DOBLANK 2 /* perform blank interpretation */
#define DOGLOB 4 /* interpret [?* */
#define DOKEY 8 /* move words with `=' to 2nd arg. list */
#define DOTRIM 16 /* trim resulting string */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define DOALL (DOSUB|DOBLANK|DOGLOB|DOKEY|DOTRIM)
Run msh through indent
2004-08-04 19:16:54 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* PROTOTYPES */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
static int newfile(char *s);
static char *findeq(char *cp);
static char *cclass(char *p, int sub);
static void initarea(void);
Allow multiple shells to be enabled.
2001-08-01 17:21:35 +00:00
extern int msh_main(int argc, char **argv);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
struct brkcon {
jmp_buf brkpt;
struct brkcon *nextlev;
};
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* redirection
*/
struct ioword {
Run msh through indent
2004-08-04 19:16:54 +00:00
short io_unit; /* unit affected */
short io_flag; /* action (below) */
char *io_name; /* file name */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define IOREAD 1 /* < */
#define IOHERE 2 /* << (here file) */
#define IOWRITE 4 /* > */
#define IOCAT 8 /* >> */
#define IOXHERE 16 /* ${}, ` in << */
#define IODUP 32 /* >&digit */
#define IOCLOSE 64 /* >&- */
Run msh through indent
2004-08-04 19:16:54 +00:00
#define IODEFAULT (-1) /* token for default IO unit */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* parsing & execution environment
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct env {
char *linep;
struct io *iobase;
struct io *iop;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
xint *errpt; /* void * */
Run msh through indent
2004-08-04 19:16:54 +00:00
int iofd;
struct env *oenv;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
} e;
/*
* flags:
* -e: quit on error
* -k: look for name=value everywhere on command line
* -n: no execution
* -t: exit after reading and executing one command
* -v: echo as read
* -x: trace
* -u: unset variables net diagnostic
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *flag;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *null; /* null value for variable */
static int intr; /* interrupt pending */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *trap[_NSIG + 1];
static char ourtrap[_NSIG + 1];
static int trapset; /* trap pending */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static int heedint; /* heed interrupt signals */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static int yynerrs; /* yacc */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static char line[LINELIM];
static char *elinep;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* other functions
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int (*inbuilt(char *s)) (struct op *);
static char *rexecve(char *c, char **v, char **envp);
static char *space(int n);
static char *strsave(char *s, int a);
static char *evalstr(char *cp, int f);
static char *putn(int n);
static char *itoa(int n);
static char *unquote(char *as);
static struct var *lookup(char *n);
static int rlookup(char *n);
static struct wdblock *glob(char *cp, struct wdblock *wb);
static int my_getc(int ec);
static int subgetc(int ec, int quoted);
Jonas Holmberg from axis dot com writes: This patch makes msh handle variable expansion within backticks more correctly. Current behaviour (wrong): -------------------------- BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` hello $ New behaviour (correct): ------------------------ BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` `echo hello` $ The current behaviour (wrong according to standards) was actually my fault. msh handles backticks by executing a subshell (which makes it work on MMU-less systems). Executing a subshell makes it hard to only expand variables once in the parent. Therefore I export all variables that will be expanded within the backticks and let the subshell handle the expansion instead. The bug was found while searching for security leaks in CGI-scripts. Current behaviour of msh makes it easy to expand backticks by mistake in $QUERY_STRING. I recommend appling the patch before release of bb 1.00. /Jonas
2004-09-02 23:13:10 +00:00
static char **makenv(int all, struct wdblock *wb);
Run msh through indent
2004-08-04 19:16:54 +00:00
static char **eval(char **ap, int f);
static int setstatus(int s);
static int waitfor(int lastpid, int canintr);
static void onintr(int s); /* SIGINT handler */
static int newenv(int f);
static void quitenv(void);
static void err(char *s);
static int anys(char *s1, char *s2);
static int any(int c, char *s);
static void next(int f);
static void setdash(void);
static void onecommand(void);
static void runtrap(int i);
static int gmatch(char *s, char *p);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* error handling
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static void leave(void); /* abort shell (or fail in subshell) */
static void fail(void); /* fail but return to process next command */
static void warn(char *s);
static void sig(int i); /* default signal handler */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* -------- area stuff -------- */
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define REGSIZE sizeof(struct region)
#define GROWBY (256)
/* #define SHRINKBY (64) */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#undef SHRINKBY
just whitespace
2006-01-25 00:08:53 +00:00
#define FREE (32767)
#define BUSY (0)
#define ALIGN (sizeof(int)-1)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct region {
Run msh through indent
2004-08-04 19:16:54 +00:00
struct region *next;
int area;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
/* -------- grammar stuff -------- */
typedef union {
Run msh through indent
2004-08-04 19:16:54 +00:00
char *cp;
char **wp;
int i;
struct op *o;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
} YYSTYPE;
Run msh through indent
2004-08-04 19:16:54 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define WORD 256
#define LOGAND 257
#define LOGOR 258
#define BREAK 259
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define IF 260
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define THEN 261
#define ELSE 262
#define ELIF 263
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define FI 264
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define CASE 265
#define ESAC 266
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define FOR 267
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define WHILE 268
#define UNTIL 269
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define DO 270
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define DONE 271
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define IN 272
/* Added for "." file expansion */
#define DOT 273
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define YYERRCODE 300
/* flags to yylex */
Run msh through indent
2004-08-04 19:16:54 +00:00
#define CONTIN 01 /* skip new lines to complete command */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define SYNTAXERR zzerr()
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *pipeline(int cf);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
static struct op *andor(void);
static struct op *c_list(void);
Run msh through indent
2004-08-04 19:16:54 +00:00
static int synio(int cf);
static void musthave(int c, int cf);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
static struct op *simple(void);
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *nested(int type, int mark);
static struct op *command(int cf);
static struct op *dogroup(int onlydone);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
static struct op *thenpart(void);
static struct op *elsepart(void);
static struct op *caselist(void);
static struct op *casepart(void);
static char **pattern(void);
static char **wordlist(void);
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *list(struct op *t1, struct op *t2);
static struct op *block(int type, struct op *t1, struct op *t2, char **wp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
static struct op *newtp(void);
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *namelist(struct op *t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
static char **copyw(void);
Run msh through indent
2004-08-04 19:16:54 +00:00
static void word(char *cp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
static struct ioword **copyio(void);
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct ioword *io(int u, int f, char *cp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
static void zzerr(void);
Run msh through indent
2004-08-04 19:16:54 +00:00
static void yyerror(char *s);
static int yylex(int cf);
static int collect(int c, int c1);
static int dual(int c);
static void diag(int ec);
static char *tree(unsigned size);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* -------- var.h -------- */
Run msh through indent
2004-08-04 19:16:54 +00:00
struct var {
char *value;
char *name;
struct var *next;
char status;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
Run msh through indent
2004-08-04 19:16:54 +00:00
#define COPYV 1 /* flag to setval, suggesting copy */
#define RONLY 01 /* variable is read-only */
#define EXPORT 02 /* variable is to be exported */
#define GETCELL 04 /* name & value space was got with getcell */
static int yyparse(void);
static struct var *lookup(char *n);
static void setval(struct var *vp, char *val);
static void nameval(struct var *vp, char *val, char *name);
static void export(struct var *vp);
static void ronly(struct var *vp);
static int isassign(char *s);
static int checkname(char *cp);
static int assign(char *s, int cf);
static void putvlist(int f, int out);
static int eqname(char *n1, char *n2);
static int execute(struct op *t, int *pin, int *pout, int act);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* -------- io.h -------- */
/* io buffer */
struct iobuf {
Run msh through indent
2004-08-04 19:16:54 +00:00
unsigned id; /* buffer id */
char buf[512]; /* buffer */
char *bufp; /* pointer into buffer */
char *ebufp; /* pointer to end of buffer */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
/* possible arguments to an IO function */
struct ioarg {
Run msh through indent
2004-08-04 19:16:54 +00:00
char *aword;
char **awordlist;
int afile; /* file descriptor */
unsigned afid; /* buffer id */
long afpos; /* file position */
struct iobuf *afbuf; /* buffer for this file */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
Run msh through indent
2004-08-04 19:16:54 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
//static struct ioarg ioargstack[NPUSH];
#define AFID_NOBUF (~0)
#define AFID_ID 0
/* an input generator's state */
Run msh through indent
2004-08-04 19:16:54 +00:00
struct io {
int (*iofn) (struct ioarg *, struct io *);
struct ioarg *argp;
int peekc;
char prev; /* previous character read by readc() */
char nlcount; /* for `'s */
char xchar; /* for `'s */
char task; /* reason for pushed IO */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
Run msh through indent
2004-08-04 19:16:54 +00:00
//static struct io iostack[NPUSH];
#define XOTHER 0 /* none of the below */
#define XDOLL 1 /* expanding ${} */
#define XGRAVE 2 /* expanding `'s */
#define XIO 3 /* file IO */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* in substitution */
#define INSUB() (e.iop->task == XGRAVE || e.iop->task == XDOLL)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* input generators for IO structure
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int nlchar(struct ioarg *ap);
static int strchar(struct ioarg *ap);
static int qstrchar(struct ioarg *ap);
static int filechar(struct ioarg *ap);
static int herechar(struct ioarg *ap);
static int linechar(struct ioarg *ap);
static int gravechar(struct ioarg *ap, struct io *iop);
static int qgravechar(struct ioarg *ap, struct io *iop);
static int dolchar(struct ioarg *ap);
static int wdchar(struct ioarg *ap);
static void scraphere(void);
static void freehere(int area);
static void gethere(void);
static void markhere(char *s, struct ioword *iop);
static int herein(char *hname, int xdoll);
static int run(struct ioarg *argp, int (*f) (struct ioarg *));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* IO functions
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int eofc(void);
static int readc(void);
static void unget(int c);
static void ioecho(int c);
data --> text
2005-10-17 09:48:57 +00:00
static void prs(const char *s);
Run msh through indent
2004-08-04 19:16:54 +00:00
static void prn(unsigned u);
static void closef(int i);
static void closeall(void);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* IO control
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static void pushio(struct ioarg *argp, int (*f) (struct ioarg *));
static int remap(int fd);
static int openpipe(int *pv);
static void closepipe(int *pv);
static struct io *setbase(struct io *ip);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define PUSHIO(what,arg,gen) ((temparg.what = (arg)),pushio(&temparg,(gen)))
#define RUN(what,arg,gen) ((temparg.what = (arg)), run(&temparg,(gen)))
/* -------- word.h -------- */
Run msh through indent
2004-08-04 19:16:54 +00:00
#define NSTART 16 /* default number of words to allow for initially */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
struct wdblock {
short w_bsize;
short w_nword;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* bounds are arbitrary */
Run msh through indent
2004-08-04 19:16:54 +00:00
char *w_words[1];
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct wdblock *addword(char *wd, struct wdblock *wb);
static struct wdblock *newword(int nw);
static char **getwords(struct wdblock *wb);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* -------- area.h -------- */
/*
* storage allocation
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *getcell(unsigned nbytes);
static void garbage(void);
static void setarea(char *cp, int a);
static int getarea(char *cp);
static void freearea(int a);
static void freecell(char *cp);
static int areanum; /* current allocation area */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#define NEW(type) (type *)getcell(sizeof(type))
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define DELETE(obj) freecell((char *)obj)
/* -------- misc stuff -------- */
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
static int forkexec(struct op *t, int *pin, int *pout, int act, char **wp);
Run msh through indent
2004-08-04 19:16:54 +00:00
static int iosetup(struct ioword *iop, int pipein, int pipeout);
static void echo(char **wp);
static struct op **find1case(struct op *t, char *w);
static struct op *findcase(struct op *t, char *w);
static void brkset(struct brkcon *bc);
static int dolabel(struct op *t);
static int dohelp(struct op *t);
static int dochdir(struct op *t);
static int doshift(struct op *t);
static int dologin(struct op *t);
static int doumask(struct op *t);
static int doexec(struct op *t);
static int dodot(struct op *t);
static int dowait(struct op *t);
static int doread(struct op *t);
static int doeval(struct op *t);
static int dotrap(struct op *t);
static int getsig(char *s);
static void setsig(int n, sighandler_t f);
static int getn(char *as);
static int dobreak(struct op *t);
static int docontinue(struct op *t);
static int brkcontin(char *cp, int val);
static int doexit(struct op *t);
static int doexport(struct op *t);
static int doreadonly(struct op *t);
static void rdexp(char **wp, void (*f) (struct var *), int key);
static void badid(char *s);
static int doset(struct op *t);
static void varput(char *s, int out);
static int dotimes(struct op *t);
static int expand(char *cp, struct wdblock **wbp, int f);
static char *blank(int f);
static int dollar(int quoted);
static int grave(int quoted);
static void globname(char *we, char *pp);
static char *generate(char *start1, char *end1, char *middle, char *end);
static int anyspcl(struct wdblock *wb);
static int xstrcmp(char *p1, char *p2);
static void glob0(char *a0, unsigned int a1, int a2,
int (*a3) (char *, char *));
static void glob1(char *base, char *lim);
static void glob2(char *i, char *j);
static void glob3(char *i, char *j, char *k);
static void readhere(char **name, char *s, int ec);
static void pushio(struct ioarg *argp, int (*f) (struct ioarg *));
static int xxchar(struct ioarg *ap);
struct here {
char *h_tag;
int h_dosub;
struct ioword *h_iop;
struct here *h_next;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
data --> text
2005-10-17 09:48:57 +00:00
static const char * const signame[] = {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
"Signal 0",
"Hangup",
Run msh through indent
2004-08-04 19:16:54 +00:00
(char *) NULL, /* interrupt */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
"Quit",
"Illegal instruction",
"Trace/BPT trap",
"Abort",
"Bus error",
"Floating Point Exception",
"Killed",
"SIGUSR1",
"SIGSEGV",
"SIGUSR2",
Run msh through indent
2004-08-04 19:16:54 +00:00
(char *) NULL, /* broken pipe */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
"Alarm clock",
"Terminated",
};
Run msh through indent
2004-08-04 19:16:54 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#define NSIGNAL (sizeof(signame)/sizeof(signame[0]))
struct res {
data --> text
2005-10-17 09:48:57 +00:00
const char *r_name;
Run msh through indent
2004-08-04 19:16:54 +00:00
int r_val;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
data --> text
2005-10-17 09:48:57 +00:00
static const struct res restab[] = {
Run msh through indent
2004-08-04 19:16:54 +00:00
{"for", FOR},
{"case", CASE},
{"esac", ESAC},
{"while", WHILE},
{"do", DO},
{"done", DONE},
{"if", IF},
{"in", IN},
{"then", THEN},
{"else", ELSE},
{"elif", ELIF},
{"until", UNTIL},
{"fi", FI},
{";;", BREAK},
{"||", LOGOR},
{"&&", LOGAND},
{"{", '{'},
{"}", '}'},
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
{".", DOT},
Run msh through indent
2004-08-04 19:16:54 +00:00
{0, 0},
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
struct builtincmd {
const char *name;
Run msh through indent
2004-08-04 19:16:54 +00:00
int (*builtinfunc) (struct op * t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
Run msh through indent
2004-08-04 19:16:54 +00:00
static const struct builtincmd builtincmds[] = {
{".", dodot},
{":", dolabel},
{"break", dobreak},
{"cd", dochdir},
{"continue", docontinue},
{"eval", doeval},
{"exec", doexec},
{"exit", doexit},
{"export", doexport},
{"help", dohelp},
{"login", dologin},
{"newgrp", dologin},
{"read", doread},
{"readonly", doreadonly},
{"set", doset},
{"shift", doshift},
{"times", dotimes},
{"trap", dotrap},
{"umask", doumask},
{"wait", dowait},
{0, 0}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
};
Reduce exported from msh applet
2005-09-22 14:38:17 +00:00
static struct op *scantree(struct op *);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
static struct op *dowholefile(int, int);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* Globals */
Run msh through indent
2004-08-04 19:16:54 +00:00
extern char **environ; /* environment pointer */
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static char **dolv;
static int dolc;
static int exstat;
static char gflg;
data --> text
2005-10-17 09:48:57 +00:00
static int interactive; /* Is this an interactive shell */
Run msh through indent
2004-08-04 19:16:54 +00:00
static int execflg;
static int multiline; /* \n changed to ; */
static struct op *outtree; /* result from parser */
static xint *failpt;
static xint *errpt;
static struct brkcon *brklist;
static int isbreak;
static struct wdblock *wdlist;
static struct wdblock *iolist;
static char *trap[_NSIG + 1];
static char ourtrap[_NSIG + 1];
static int trapset; /* trap pending */
static int yynerrs; /* yacc */
static char line[LINELIM];
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#ifdef MSHDEBUG
static struct var *mshdbg_var;
#endif
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct var *vlist; /* dictionary */
static struct var *homedir; /* home directory */
static struct var *prompt; /* main prompt */
static struct var *cprompt; /* continuation prompt */
static struct var *path; /* search path for commands */
static struct var *shell; /* shell to interpret command files */
static struct var *ifs; /* field separators */
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static int areanum; /* current allocation area */
static int intr;
static int inparse;
static char flags['z' - 'a' + 1];
static char *flag = flags - 'a';
static char *null = "";
static int heedint = 1;
static void (*qflag) (int) = SIG_IGN;
static int startl;
static int peeksym;
static int nlseen;
static int iounit = IODEFAULT;
static YYSTYPE yylval;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
static char *elinep = line + sizeof(line) - 5;
static struct ioarg temparg = { 0, 0, 0, AFID_NOBUF, 0 }; /* temporary for PUSHIO */
static struct ioarg ioargstack[NPUSH];
static struct io iostack[NPUSH];
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct iobuf sharedbuf = { AFID_NOBUF };
static struct iobuf mainbuf = { AFID_NOBUF };
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
static unsigned bufid = AFID_ID; /* buffer id counter */
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct here *inhere; /* list of hear docs while parsing */
static struct here *acthere; /* list of active here documents */
static struct region *areabot; /* bottom of area */
static struct region *areatop; /* top of area */
static struct region *areanxt; /* starting point of scan */
static void *brktop;
static void *brkaddr;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
static struct env e = {
line, /* linep: char ptr */
iostack, /* iobase: struct io ptr */
iostack - 1, /* iop: struct io ptr */
(xint *) NULL, /* errpt: void ptr for errors? */
FDBASE, /* iofd: file desc */
(struct env *) NULL /* oenv: struct env ptr */
};
#ifdef MSHDEBUG
void print_t(struct op *t)
{
DBGPRINTF(("T: t=0x%x, type %s, words=0x%x, IOword=0x%x\n", t,
T_CMD_NAMES[t->type], t->words, t->ioact));
if (t->words) {
DBGPRINTF(("T: W1: %s", t->words[0]));
}
return;
}
void print_tree(struct op *head)
{
if (head == NULL) {
DBGPRINTF(("PRINT_TREE: no tree\n"));
return;
}
DBGPRINTF(("NODE: 0x%x, left 0x%x, right 0x%x\n", head, head->left,
head->right));
if (head->left)
print_tree(head->left);
if (head->right)
print_tree(head->right);
return;
}
#endif /* MSHDEBUG */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Major rework of the directory structure and the entire build system. -Erik
2001-10-24 05:00:29 +00:00
#ifdef CONFIG_FEATURE_COMMAND_EDITING
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *current_prompt;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#endif
/* -------- sh.c -------- */
/*
* shell
*/
Robert P. Day removed 8 gazillion occurrences of "extern" on function definitions. (That should only be on prototypes.)
2006-03-06 20:47:33 +00:00
int msh_main(int argc, char **argv)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int f;
REGISTER char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int cflag;
char *name, **ap;
Run msh through indent
2004-08-04 19:16:54 +00:00
int (*iof) (struct ioarg *);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("MSH_MAIN: argc %d, environ 0x%x\n", argc, environ));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
initarea();
if ((ap = environ) != NULL) {
while (*ap)
assign(*ap++, !COPYV);
for (ap = environ; *ap;)
export(lookup(*ap++));
}
closeall();
areanum = 1;
shell = lookup("SHELL");
if (shell->value == null)
Quiet a few warnings
2004-08-27 19:55:28 +00:00
setval(shell, (char *)DEFAULT_SHELL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
export(shell);
homedir = lookup("HOME");
if (homedir->value == null)
setval(homedir, "/");
export(homedir);
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
setval(lookup("$"), putn(getpid()));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
path = lookup("PATH");
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
if (path->value == null) {
if (geteuid() == 0)
setval(path, "/sbin:/bin:/usr/sbin:/usr/bin");
else
setval(path, "/bin:/usr/bin");
}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
export(path);
ifs = lookup("IFS");
if (ifs->value == null)
setval(ifs, " \t\n");
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#ifdef MSHDEBUG
mshdbg_var = lookup("MSHDEBUG");
if (mshdbg_var->value == null)
setval(mshdbg_var, "0");
#endif
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
prompt = lookup("PS1");
Major rework of the directory structure and the entire build system. -Erik
2001-10-24 05:00:29 +00:00
#ifdef CONFIG_FEATURE_SH_FANCY_PROMPT
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (prompt->value == null)
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
#endif
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
setval(prompt, "$ ");
if (geteuid() == 0) {
setval(prompt, "# ");
prompt->status &= ~EXPORT;
}
cprompt = lookup("PS2");
Major rework of the directory structure and the entire build system. -Erik
2001-10-24 05:00:29 +00:00
#ifdef CONFIG_FEATURE_SH_FANCY_PROMPT
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (cprompt->value == null)
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
#endif
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
setval(cprompt, "> ");
iof = filechar;
cflag = 0;
name = *argv++;
if (--argc >= 1) {
Run msh through indent
2004-08-04 19:16:54 +00:00
if (argv[0][0] == '-' && argv[0][1] != '\0') {
for (s = argv[0] + 1; *s; s++)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
switch (*s) {
case 'c':
prompt->status &= ~EXPORT;
cprompt->status &= ~EXPORT;
setval(prompt, "");
setval(cprompt, "");
cflag = 1;
if (--argc > 0)
PUSHIO(aword, *++argv, iof = nlchar);
break;
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case 'q':
qflag = SIG_DFL;
break;
case 's':
/* standard input */
break;
case 't':
prompt->status &= ~EXPORT;
setval(prompt, "");
iof = linechar;
break;
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case 'i':
interactive++;
default:
Run msh through indent
2004-08-04 19:16:54 +00:00
if (*s >= 'a' && *s <= 'z')
flag[(int) *s]++;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
} else {
argv--;
argc++;
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (iof == filechar && --argc > 0) {
setval(prompt, "");
setval(cprompt, "");
prompt->status &= ~EXPORT;
cprompt->status &= ~EXPORT;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* Shell is non-interactive, activate printf-based debug */
#ifdef MSHDEBUG
mshdbg = (int) (((char) (mshdbg_var->value[0])) - '0');
if (mshdbg < 0)
mshdbg = 0;
#endif
DBGPRINTF(("MSH_MAIN: calling newfile()\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (newfile(name = *++argv))
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
exit(1); /* Exit on error */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
setdash();
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* This won't be true if PUSHIO has been called, say from newfile() above */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (e.iop < iostack) {
PUSHIO(afile, 0, iof);
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
if (isatty(0) && isatty(1) && !cflag) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
interactive++;
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
#ifndef CONFIG_FEATURE_SH_EXTRA_QUIET
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#ifdef MSHDEBUG
moved BB_BANNER to applets/version.c file: make kernel like version, removed depend loop: busybox.h depend with BB_BT, and all sources depend with busybox.h
2006-02-16 15:40:24 +00:00
printf("\n\n%s Built-in shell (msh with debug)\n", BB_BANNER);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#else
moved BB_BANNER to applets/version.c file: make kernel like version, removed depend loop: busybox.h depend with BB_BT, and all sources depend with busybox.h
2006-02-16 15:40:24 +00:00
printf("\n\n%s Built-in shell (msh)\n", BB_BANNER);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#endif
Run msh through indent
2004-08-04 19:16:54 +00:00
printf("Enter 'help' for a list of built-in commands.\n\n");
Add an option to make the shells not advertise their busybox nature
2001-10-19 00:22:23 +00:00
#endif
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
signal(SIGQUIT, qflag);
if (name && name[0] == '-') {
interactive++;
if ((f = open(".profile", 0)) >= 0)
next(remap(f));
if ((f = open("/etc/profile", 0)) >= 0)
next(remap(f));
}
if (interactive)
signal(SIGTERM, sig);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (signal(SIGINT, SIG_IGN) != SIG_IGN)
signal(SIGINT, onintr);
dolv = argv;
dolc = argc;
dolv[0] = name;
if (dolc > 1) {
for (ap = ++argv; --argc > 0;) {
if (assign(*ap = *argv++, !COPYV)) {
Run msh through indent
2004-08-04 19:16:54 +00:00
dolc--; /* keyword */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
} else {
ap++;
}
}
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
setval(lookup("#"), putn((--dolc < 0) ? (dolc = 0) : dolc));
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("MSH_MAIN: begin FOR loop, interactive %d, e.iop 0x%x, iostack 0x%x\n", interactive, e.iop, iostack));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (;;) {
if (interactive && e.iop <= iostack) {
Major rework of the directory structure and the entire build system. -Erik
2001-10-24 05:00:29 +00:00
#ifdef CONFIG_FEATURE_COMMAND_EDITING
Run msh through indent
2004-08-04 19:16:54 +00:00
current_prompt = prompt->value;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#else
prs(prompt->value);
#endif
}
onecommand();
Ensure that getenv("PATH") stays current, since otherwise cmdedit (which calls getenv("PATH")) would not operate upon the current PATH settings, which was really quite iritating. -Erik
2002-04-26 23:40:09 +00:00
/* Ensure that getenv("PATH") stays current */
setenv("PATH", path->value, 1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("MSH_MAIN: returning.\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void setdash()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp;
REGISTER int c;
Run msh through indent
2004-08-04 19:16:54 +00:00
char m['z' - 'a' + 1];
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
cp = m;
Run msh through indent
2004-08-04 19:16:54 +00:00
for (c = 'a'; c <= 'z'; c++)
if (flag[(int) c])
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
*cp++ = c;
*cp = 0;
setval(lookup("-"), m);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int newfile(s)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int f;
DBGPRINTF7(("NEWFILE: opening %s\n", s));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (strcmp(s, "-") != 0) {
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("NEWFILE: s is %s\n", s));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
f = open(s, 0);
if (f < 0) {
prs(s);
err(": cannot open");
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
} else
f = 0;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
next(remap(f));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
struct op *scantree(head)
struct op *head;
{
struct op *dotnode;
if (head == NULL)
return (NULL);
if (head->left != NULL) {
dotnode = scantree(head->left);
if (dotnode)
return (dotnode);
}
if (head->right != NULL) {
dotnode = scantree(head->right);
if (dotnode)
return (dotnode);
}
if (head->words == NULL)
return (NULL);
DBGPRINTF5(("SCANTREE: checking node 0x%x\n", head));
if ((head->type != TDOT) && (strcmp(".", head->words[0]) == 0)) {
DBGPRINTF5(("SCANTREE: dot found in node 0x%x\n", head));
return (head);
}
return (NULL);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void onecommand()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
jmp_buf m1;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("ONECOMMAND: enter, outtree=0x%x\n", outtree));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
while (e.oenv)
quitenv();
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
areanum = 1;
freehere(areanum);
freearea(areanum);
garbage();
wdlist = 0;
iolist = 0;
e.errpt = 0;
e.linep = line;
yynerrs = 0;
multiline = 0;
inparse = 1;
intr = 0;
execflg = 0;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
setjmp(failpt = m1); /* Bruce Evans' fix */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (setjmp(failpt = m1) || yyparse() || intr) {
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("ONECOMMAND: this is not good.\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
while (e.oenv)
quitenv();
scraphere();
if (!interactive && intr)
leave();
inparse = 0;
intr = 0;
return;
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
inparse = 0;
brklist = 0;
intr = 0;
execflg = 0;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if (!flag['n']) {
DBGPRINTF(("ONECOMMAND: calling execute, t=outtree=0x%x\n",
outtree));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
execute(outtree, NOPIPE, NOPIPE, 0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (!interactive && intr) {
execflg = 0;
leave();
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((i = trapset) != 0) {
trapset = 0;
runtrap(i);
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void fail()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
longjmp(failpt, 1);
/* NOTREACHED */
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void leave()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("LEAVE: leave called!\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (execflg)
fail();
scraphere();
freehere(1);
runtrap(0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
_exit(exstat);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* NOTREACHED */
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void warn(s)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Run msh through indent
2004-08-04 19:16:54 +00:00
if (*s) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
prs(s);
exstat = -1;
}
prs("\n");
if (flag['e'])
leave();
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void err(s)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *s;
{
warn(s);
if (flag['n'])
return;
if (!interactive)
leave();
if (e.errpt)
longjmp(e.errpt, 1);
closeall();
e.iop = e.iobase = iostack;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int newenv(f)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int f;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct env *ep;
DBGPRINTF(("NEWENV: f=%d (indicates quitenv and return)\n", f));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (f) {
quitenv();
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
ep = (struct env *) space(sizeof(*ep));
if (ep == NULL) {
while (e.oenv)
quitenv();
fail();
}
*ep = e;
e.oenv = ep;
e.errpt = errpt;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void quitenv()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct env *ep;
REGISTER int fd;
DBGPRINTF(("QUITENV: e.oenv=0x%x\n", e.oenv));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((ep = e.oenv) != NULL) {
fd = e.iofd;
e = *ep;
/* should close `'d files */
DELETE(ep);
while (--fd >= e.iofd)
close(fd);
}
}
/*
* Is any character from s1 in s2?
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int anys(s1, s2)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s1, *s2;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
while (*s1)
if (any(*s1++, s2))
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Is character c in s?
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int any(c, s)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
REGISTER char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
while (*s)
if (*s++ == c)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *putn(n)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int n;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Run msh through indent
2004-08-04 19:16:54 +00:00
return (itoa(n));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *itoa(n)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int n;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
static char s[20];
Run msh through indent
2004-08-04 19:16:54 +00:00
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
snprintf(s, sizeof(s), "%u", n);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (s);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static void next(int f)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
PUSHIO(afile, f, filechar);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void onintr(s)
int s; /* ANSI C requires a parameter */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
signal(SIGINT, onintr);
intr = 1;
if (interactive) {
if (inparse) {
prs("\n");
fail();
}
Run msh through indent
2004-08-04 19:16:54 +00:00
} else if (heedint) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
execflg = 0;
leave();
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *space(n)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int n;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((cp = getcell(n)) == 0)
err("out of string space");
Run msh through indent
2004-08-04 19:16:54 +00:00
return (cp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *strsave(s, a)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int a;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp, *xp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((cp = space(strlen(s) + 1)) != NULL) {
setarea((char *) cp, a);
for (xp = cp; (*xp++ = *s++) != '\0';);
return (cp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return ("");
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* trap handling
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static void sig(i)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
trapset = i;
signal(i, sig);
}
static void runtrap(i)
int i;
{
char *trapstr;
if ((trapstr = trap[i]) == NULL)
return;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (i == 0)
trap[i] = 0;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
RUN(aword, trapstr, nlchar);
}
/* -------- var.c -------- */
/*
* Find the given name in the dictionary
* and return its value. If the name was
* not previously there, enter it now and
* return a null value.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct var *lookup(n)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *n;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct var *vp;
REGISTER char *cp;
REGISTER int c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
static struct var dummy;
if (isdigit(*n)) {
dummy.name = n;
for (c = 0; isdigit(*n) && c < 1000; n++)
Run msh through indent
2004-08-04 19:16:54 +00:00
c = c * 10 + *n - '0';
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
dummy.status = RONLY;
Run msh through indent
2004-08-04 19:16:54 +00:00
dummy.value = c <= dolc ? dolv[c] : null;
return (&dummy);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
for (vp = vlist; vp; vp = vp->next)
if (eqname(vp->name, n))
Run msh through indent
2004-08-04 19:16:54 +00:00
return (vp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
cp = findeq(n);
Run msh through indent
2004-08-04 19:16:54 +00:00
vp = (struct var *) space(sizeof(*vp));
if (vp == 0 || (vp->name = space((int) (cp - n) + 2)) == 0) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
dummy.name = dummy.value = "";
Run msh through indent
2004-08-04 19:16:54 +00:00
return (&dummy);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
for (cp = vp->name; (*cp = *n++) && *cp != '='; cp++);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (*cp == 0)
*cp = '=';
*++cp = 0;
Run msh through indent
2004-08-04 19:16:54 +00:00
setarea((char *) vp, 0);
setarea((char *) vp->name, 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
vp->value = null;
vp->next = vlist;
vp->status = GETCELL;
vlist = vp;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (vp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* give variable at `vp' the value `val'.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static void setval(vp, val)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct var *vp;
char *val;
{
Run msh through indent
2004-08-04 19:16:54 +00:00
nameval(vp, val, (char *) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* if name is not NULL, it must be
* a prefix of the space `val',
* and end with `='.
* this is all so that exporting
* values is reasonably painless.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static void nameval(vp, val, name)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct var *vp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *val, *name;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp, *xp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *nv;
int fl;
if (vp->status & RONLY) {
for (xp = vp->name; *xp && *xp != '=';)
putc(*xp++, stderr);
err(" is read-only");
return;
}
fl = 0;
if (name == NULL) {
Run msh through indent
2004-08-04 19:16:54 +00:00
xp = space(strlen(vp->name) + strlen(val) + 2);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (xp == 0)
return;
/* make string: name=value */
Run msh through indent
2004-08-04 19:16:54 +00:00
setarea((char *) xp, 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
name = xp;
Run msh through indent
2004-08-04 19:16:54 +00:00
for (cp = vp->name; (*xp = *cp++) && *xp != '='; xp++);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (*xp++ == 0)
xp[-1] = '=';
nv = xp;
Run msh through indent
2004-08-04 19:16:54 +00:00
for (cp = val; (*xp++ = *cp++) != '\0';);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
val = nv;
fl = GETCELL;
}
if (vp->status & GETCELL)
Run msh through indent
2004-08-04 19:16:54 +00:00
freecell(vp->name); /* form new string `name=value' */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
vp->name = name;
vp->value = val;
vp->status |= fl;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void export(vp)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct var *vp;
{
vp->status |= EXPORT;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void ronly(vp)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct var *vp;
{
Add support for underscores in variable names.
2001-08-06 14:14:18 +00:00
if (isalpha(vp->name[0]) || vp->name[0] == '_') /* not an internal symbol */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
vp->status |= RONLY;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int isassign(s)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("ISASSIGN: enter, s=%s\n", s));
Run msh through indent
2004-08-04 19:16:54 +00:00
if (!isalpha((int) *s) && *s != '_')
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (; *s != '='; s++)
Add support for underscores in variable names.
2001-08-06 14:14:18 +00:00
if (*s == 0 || (!isalnum(*s) && *s != '_'))
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int assign(s, cf)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int cf;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct var *vp;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("ASSIGN: enter, s=%s, cf=%d\n", s, cf));
Add support for underscores in variable names.
2001-08-06 14:14:18 +00:00
if (!isalpha(*s) && *s != '_')
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (cp = s; *cp != '='; cp++)
Add support for underscores in variable names.
2001-08-06 14:14:18 +00:00
if (*cp == 0 || (!isalnum(*cp) && *cp != '_'))
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
vp = lookup(s);
Run msh through indent
2004-08-04 19:16:54 +00:00
nameval(vp, ++cp, cf == COPYV ? (char *) NULL : s);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (cf != COPYV)
vp->status &= ~GETCELL;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int checkname(cp)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("CHECKNAME: enter, cp=%s\n", cp));
Run msh through indent
2004-08-04 19:16:54 +00:00
if (!isalpha(*cp++) && *(cp - 1) != '_')
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
while (*cp)
Run msh through indent
2004-08-04 19:16:54 +00:00
if (!isalnum(*cp++) && *(cp - 1) != '_')
return (0);
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void putvlist(f, out)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int f, out;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct var *vp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (vp = vlist; vp; vp = vp->next)
Add support for underscores in variable names.
2001-08-06 14:14:18 +00:00
if (vp->status & f && (isalpha(*vp->name) || *vp->name == '_')) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (vp->status & EXPORT)
write(out, "export ", 7);
if (vp->status & RONLY)
write(out, "readonly ", 9);
Run msh through indent
2004-08-04 19:16:54 +00:00
write(out, vp->name, (int) (findeq(vp->name) - vp->name));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
write(out, "\n", 1);
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int eqname(n1, n2)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *n1, *n2;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
for (; *n1 != '=' && *n1 != 0; n1++)
if (*n2++ != *n1)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
return (*n2 == 0 || *n2 == '=');
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *findeq(cp)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
while (*cp != '\0' && *cp != '=')
cp++;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (cp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/* -------- gmatch.c -------- */
/*
* int gmatch(string, pattern)
* char *string, *pattern;
*
* Match a pattern as in sh(1).
*/
#define CMASK 0377
#define QUOTE 0200
#define QMASK (CMASK&~QUOTE)
Run msh through indent
2004-08-04 19:16:54 +00:00
#define NOT '!' /* might use ^ */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static int gmatch(s, p)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s, *p;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int sc, pc;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (s == NULL || p == NULL)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
while ((pc = *p++ & CMASK) != '\0') {
sc = *s++ & QMASK;
switch (pc) {
case '[':
if ((p = cclass(p, sc)) == NULL)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
break;
case '?':
if (sc == 0)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
break;
case '*':
s--;
do {
if (*p == '\0' || gmatch(s, p))
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
} while (*s++ != '\0');
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
default:
Run msh through indent
2004-08-04 19:16:54 +00:00
if (sc != (pc & ~QUOTE))
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (*s == 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *cclass(p, sub)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *p;
REGISTER int sub;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c, d, not, found;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((not = *p == NOT) != 0)
p++;
found = not;
do {
if (*p == '\0')
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((char *) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
c = *p & CMASK;
if (p[1] == '-' && p[2] != ']') {
d = p[2] & CMASK;
p++;
} else
d = c;
if (c == sub || (c <= sub && sub <= d))
found = !not;
} while (*++p != ']');
Run msh through indent
2004-08-04 19:16:54 +00:00
return (found ? p + 1 : (char *) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/* -------- area.c -------- */
/*
* All memory between (char *)areabot and (char *)(areatop+1) is
* exclusively administered by the area management routines.
* It is assumed that sbrk() and brk() manipulate the high end.
*/
#define sbrk(X) ({ void * __q = (void *)-1; if (brkaddr + (int)(X) < brktop) { __q = brkaddr; brkaddr+=(int)(X); } __q;})
Run msh through indent
2004-08-04 19:16:54 +00:00
static void initarea()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
brkaddr = malloc(AREASIZE);
brktop = brkaddr + AREASIZE;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
fix up yet more annoying signed/unsigned and mixed type errors
2006-01-30 22:48:39 +00:00
while ((long) sbrk(0) & ALIGN)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
sbrk(1);
Run msh through indent
2004-08-04 19:16:54 +00:00
areabot = (struct region *) sbrk(REGSIZE);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
areabot->next = areabot;
areabot->area = BUSY;
areatop = areabot;
areanxt = areabot;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
char *getcell(nbytes)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
unsigned nbytes;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int nregio;
REGISTER struct region *p, *q;
REGISTER int i;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (nbytes == 0) {
puts("getcell(0)");
abort();
Run msh through indent
2004-08-04 19:16:54 +00:00
}
/* silly and defeats the algorithm */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* round upwards and add administration area
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
nregio = (nbytes + (REGSIZE - 1)) / REGSIZE + 1;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (p = areanxt;;) {
if (p->area > areanum) {
/*
* merge free cells
*/
while ((q = p->next)->area > areanum && q != areanxt)
p->next = q->next;
/*
* exit loop if cell big enough
*/
if (q >= p + nregio)
goto found;
}
p = p->next;
if (p == areanxt)
break;
}
i = nregio >= GROWBY ? nregio : GROWBY;
Run msh through indent
2004-08-04 19:16:54 +00:00
p = (struct region *) sbrk(i * REGSIZE);
if (p == (struct region *) -1)
return ((char *) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
p--;
if (p != areatop) {
puts("not contig");
Run msh through indent
2004-08-04 19:16:54 +00:00
abort(); /* allocated areas are contiguous */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
q = p + i;
p->next = q;
p->area = FREE;
q->next = areabot;
q->area = BUSY;
areatop = q;
Run msh through indent
2004-08-04 19:16:54 +00:00
found:
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* we found a FREE area big enough, pointed to by 'p', and up to 'q'
*/
areanxt = p + nregio;
if (areanxt < q) {
/*
* split into requested area and rest
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
if (areanxt + 1 > q) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
puts("OOM");
Run msh through indent
2004-08-04 19:16:54 +00:00
abort(); /* insufficient space left for admin */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
areanxt->next = q;
areanxt->area = FREE;
p->next = areanxt;
}
p->area = areanum;
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((char *) (p + 1));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void freecell(cp)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *cp;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct region *p;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((p = (struct region *) cp) != NULL) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
p--;
if (p < areanxt)
areanxt = p;
p->area = FREE;
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void freearea(a)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int a;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct region *p, *top;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
top = areatop;
for (p = areabot; p != top; p = p->next)
if (p->area >= a)
p->area = FREE;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void setarea(cp, a)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *cp;
int a;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct region *p;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((p = (struct region *) cp) != NULL)
(p - 1)->area = a;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
int getarea(cp)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *cp;
{
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((struct region *) cp - 1)->area;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void garbage()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct region *p, *q, *top;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
top = areatop;
for (p = areabot; p != top; p = p->next) {
if (p->area > areanum) {
while ((q = p->next)->area > areanum)
p->next = q->next;
areanxt = p;
}
}
#ifdef SHRINKBY
if (areatop >= q + SHRINKBY && q->area > areanum) {
Run msh through indent
2004-08-04 19:16:54 +00:00
brk((char *) (q + 1));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
q->next = areabot;
q->area = BUSY;
areatop = q;
}
#endif
}
/* -------- csyn.c -------- */
/*
* shell: syntax (C version)
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
int yyparse()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("YYPARSE: enter...\n"));
Run msh through indent
2004-08-04 19:16:54 +00:00
startl = 1;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
peeksym = 0;
yynerrs = 0;
outtree = c_list();
musthave('\n', 0);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (yynerrs != 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *pipeline(cf)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int cf;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t, *p;
REGISTER int c;
DBGPRINTF7(("PIPELINE: enter, cf=%d\n", cf));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = command(cf);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF9(("PIPELINE: t=0x%x\n", t));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (t != NULL) {
while ((c = yylex(0)) == '|') {
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if ((p = command(CONTIN)) == NULL) {
DBGPRINTF8(("PIPELINE: error!\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
SYNTAXERR;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (t->type != TPAREN && t->type != TCOM) {
/* shell statement */
t = block(TPAREN, t, NOBLOCK, NOWORDS);
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = block(TPIPE, t, p, NOWORDS);
}
peeksym = c;
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("PIPELINE: returning t=0x%x\n", t));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *andor()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t, *p;
REGISTER int c;
DBGPRINTF7(("ANDOR: enter...\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = pipeline(0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF9(("ANDOR: t=0x%x\n", t));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (t != NULL) {
while ((c = yylex(0)) == LOGAND || c == LOGOR) {
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if ((p = pipeline(CONTIN)) == NULL) {
DBGPRINTF8(("ANDOR: error!\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
SYNTAXERR;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
t = block(c == LOGAND ? TAND : TOR, t, p, NOWORDS);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
} /* WHILE */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
peeksym = c;
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("ANDOR: returning t=0x%x\n", t));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *c_list()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t, *p;
REGISTER int c;
DBGPRINTF7(("C_LIST: enter...\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = andor();
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (t != NULL) {
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((peeksym = yylex(0)) == '&')
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = block(TASYNC, t, NOBLOCK, NOWORDS);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
while ((c = yylex(0)) == ';' || c == '&'
|| (multiline && c == '\n')) {
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((p = andor()) == NULL)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((peeksym = yylex(0)) == '&')
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
p = block(TASYNC, p, NOBLOCK, NOWORDS);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = list(t, p);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
} /* WHILE */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
peeksym = c;
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* IF */
DBGPRINTF7(("C_LIST: returning t=0x%x\n", t));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int synio(cf)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int cf;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioword *iop;
REGISTER int i;
REGISTER int c;
DBGPRINTF7(("SYNIO: enter, cf=%d\n", cf));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((c = yylex(cf)) != '<' && c != '>') {
peeksym = c;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
i = yylval.i;
musthave(WORD, 0);
iop = io(iounit, i, yylval.cp);
iounit = IODEFAULT;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (i & IOHERE)
markhere(yylval.cp, iop);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("SYNIO: returning 1\n"));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void musthave(c, cf)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int c, cf;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if ((peeksym = yylex(cf)) != c) {
DBGPRINTF7(("MUSTHAVE: error!\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
SYNTAXERR;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
peeksym = 0;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *simple()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = NULL;
for (;;) {
switch (peeksym = yylex(0)) {
case '<':
case '>':
(void) synio(0);
break;
case WORD:
if (t == NULL) {
t = newtp();
t->type = TCOM;
}
peeksym = 0;
word(yylval.cp);
break;
default:
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *nested(type, mark)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int type, mark;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
DBGPRINTF3(("NESTED: enter, type=%d, mark=%d\n", type, mark));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
multiline++;
t = c_list();
musthave(mark, 0);
multiline--;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (block(type, t, NOBLOCK, NOWORDS));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *command(cf)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int cf;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct wdblock *iosave;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
DBGPRINTF(("COMMAND: enter, cf=%d\n", cf));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
iosave = iolist;
iolist = NULL;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (multiline)
cf |= CONTIN;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
while (synio(cf))
cf = 0;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
c = yylex(cf);
switch (c) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
default:
peeksym = c;
if ((t = simple()) == NULL) {
if (iolist == NULL)
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((struct op *) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = newtp();
t->type = TCOM;
}
break;
case '(':
t = nested(TPAREN, ')');
break;
case '{':
t = nested(TBRACE, '}');
break;
case FOR:
t = newtp();
t->type = TFOR;
musthave(WORD, 0);
startl = 1;
t->str = yylval.cp;
multiline++;
t->words = wordlist();
if ((c = yylex(0)) != '\n' && c != ';')
peeksym = c;
t->left = dogroup(0);
multiline--;
break;
case WHILE:
case UNTIL:
multiline++;
t = newtp();
Run msh through indent
2004-08-04 19:16:54 +00:00
t->type = c == WHILE ? TWHILE : TUNTIL;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t->left = c_list();
t->right = dogroup(1);
t->words = NULL;
multiline--;
break;
case CASE:
t = newtp();
t->type = TCASE;
musthave(WORD, 0);
t->str = yylval.cp;
startl++;
multiline++;
musthave(IN, CONTIN);
startl++;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t->left = caselist();
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
musthave(ESAC, 0);
multiline--;
break;
case IF:
multiline++;
t = newtp();
t->type = TIF;
t->left = c_list();
t->right = thenpart();
musthave(FI, 0);
multiline--;
break;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
case DOT:
t = newtp();
t->type = TDOT;
musthave(WORD, 0); /* gets name of file */
DBGPRINTF7(("COMMAND: DOT clause, yylval.cp is %s\n", yylval.cp));
word(yylval.cp); /* add word to wdlist */
word(NOWORD); /* terminate wdlist */
t->words = copyw(); /* dup wdlist */
break;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
while (synio(0));
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = namelist(t);
iolist = iosave;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("COMMAND: returning 0x%x\n", t));
return (t);
}
static struct op *dowholefile(type, mark)
int type;
int mark;
{
REGISTER struct op *t;
DBGPRINTF(("DOWHOLEFILE: enter, type=%d, mark=%d\n", type, mark));
multiline++;
t = c_list();
multiline--;
t = block(type, t, NOBLOCK, NOWORDS);
DBGPRINTF(("DOWHOLEFILE: return t=0x%x\n", t));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *dogroup(onlydone)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int onlydone;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
REGISTER struct op *mylist;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
c = yylex(CONTIN);
if (c == DONE && onlydone)
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((struct op *) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (c != DO)
SYNTAXERR;
mylist = c_list();
musthave(DONE, 0);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (mylist);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *thenpart()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((c = yylex(0)) != THEN) {
peeksym = c;
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((struct op *) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
t = newtp();
t->type = 0;
t->left = c_list();
if (t->left == NULL)
SYNTAXERR;
t->right = elsepart();
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *elsepart()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
switch (c = yylex(0)) {
case ELSE:
if ((t = c_list()) == NULL)
SYNTAXERR;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case ELIF:
t = newtp();
t->type = TELIF;
t->left = c_list();
t->right = thenpart();
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
default:
peeksym = c;
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((struct op *) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *caselist()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = NULL;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
while ((peeksym = yylex(CONTIN)) != ESAC) {
DBGPRINTF(("CASELIST, doing yylex, peeksym=%d\n", peeksym));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = list(t, casepart());
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
DBGPRINTF(("CASELIST, returning t=0x%x\n", t));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *casepart()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
DBGPRINTF7(("CASEPART: enter...\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = newtp();
t->type = TPAT;
t->words = pattern();
musthave(')', 0);
t->left = c_list();
if ((peeksym = yylex(CONTIN)) != ESAC)
musthave(BREAK, CONTIN);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("CASEPART: made newtp(TPAT, t=0x%x)\n", t));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static char **pattern()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c, cf;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
cf = CONTIN;
do {
musthave(WORD, cf);
word(yylval.cp);
cf = 0;
} while ((c = yylex(0)) == '|');
peeksym = c;
word(NOWORD);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
return (copyw());
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static char **wordlist()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((c = yylex(0)) != IN) {
peeksym = c;
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((char **) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
startl = 0;
while ((c = yylex(0)) == WORD)
word(yylval.cp);
word(NOWORD);
peeksym = c;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (copyw());
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* supporting functions
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *list(t1, t2)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t1, *t2;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("LIST: enter, t1=0x%x, t2=0x%x\n", t1, t2));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (t1 == NULL)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t2);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (t2 == NULL)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t1);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
return (block(TLIST, t1, t2, NOWORDS));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *block(type, t1, t2, wp)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int type;
struct op *t1, *t2;
char **wp;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
DBGPRINTF7(("BLOCK: enter, type=%d (%s)\n", type, T_CMD_NAMES[type]));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t = newtp();
t->type = type;
t->left = t1;
t->right = t2;
t->words = wp;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("BLOCK: inserted 0x%x between 0x%x and 0x%x\n", t, t1,
t2));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* See if given string is a shell multiline (FOR, IF, etc) */
Run msh through indent
2004-08-04 19:16:54 +00:00
static int rlookup(n)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *n;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
data --> text
2005-10-17 09:48:57 +00:00
REGISTER const struct res *rp;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("RLOOKUP: enter, n is %s\n", n));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (rp = restab; rp->r_name; rp++)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if (strcmp(rp->r_name, n) == 0) {
DBGPRINTF7(("RLOOKUP: match, returning %d\n", rp->r_val));
return (rp->r_val); /* Return numeric code for shell multiline */
}
DBGPRINTF7(("RLOOKUP: NO match, returning 0\n"));
return (0); /* Not a shell multiline */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *newtp()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
t = (struct op *) tree(sizeof(*t));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t->type = 0;
t->words = NULL;
t->ioact = NULL;
t->left = NULL;
t->right = NULL;
t->str = NULL;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF3(("NEWTP: allocated 0x%x\n", t));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *namelist(t)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("NAMELIST: enter, t=0x%x, type %s, iolist=0x%x\n", t,
T_CMD_NAMES[t->type], iolist));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (iolist) {
Run msh through indent
2004-08-04 19:16:54 +00:00
iolist = addword((char *) NULL, iolist);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t->ioact = copyio();
} else
t->ioact = NULL;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (t->type != TCOM) {
if (t->type != TPAREN && t->ioact != NULL) {
t = block(TPAREN, t, NOBLOCK, NOWORDS);
t->ioact = t->left->ioact;
t->left->ioact = NULL;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
word(NOWORD);
t->words = copyw();
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static char **copyw()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char **wd;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
wd = getwords(wdlist);
wdlist = 0;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (wd);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void word(cp)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *cp;
{
wdlist = addword(cp, wdlist);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct ioword **copyio()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioword **iop;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
iop = (struct ioword **) getwords(iolist);
iolist = 0;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (iop);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct ioword *io(u, f, cp)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int u;
int f;
char *cp;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioword *iop;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
iop = (struct ioword *) tree(sizeof(*iop));
iop->io_unit = u;
iop->io_flag = f;
iop->io_name = cp;
Run msh through indent
2004-08-04 19:16:54 +00:00
iolist = addword((char *) iop, iolist);
return (iop);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void zzerr()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
yyerror("syntax error");
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void yyerror(s)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *s;
{
yynerrs++;
if (interactive && e.iop <= iostack) {
multiline = 0;
Run msh through indent
2004-08-04 19:16:54 +00:00
while (eofc() == 0 && yylex(0) != '\n');
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
err(s);
fail();
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int yylex(cf)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int cf;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c, c1;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int atstart;
if ((c = peeksym) > 0) {
peeksym = 0;
if (c == '\n')
startl = 1;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
nlseen = 0;
atstart = startl;
startl = 0;
yylval.i = 0;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
e.linep = line;
/* MALAMO */
line[LINELIM - 1] = '\0';
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
loop:
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
while ((c = my_getc(0)) == ' ' || c == '\t') /* Skip whitespace */
;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
switch (c) {
default:
if (any(c, "0123456789")) {
unget(c1 = my_getc(0));
if (c1 == '<' || c1 == '>') {
iounit = c - '0';
goto loop;
}
*e.linep++ = c;
c = c1;
}
break;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
case '#': /* Comment, skip to next newline or End-of-string */
Run msh through indent
2004-08-04 19:16:54 +00:00
while ((c = my_getc(0)) != 0 && c != '\n');
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
unget(c);
goto loop;
case 0:
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF5(("YYLEX: return 0, c=%d\n", c));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case '$':
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF9(("YYLEX: found $\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
*e.linep++ = c;
if ((c = my_getc(0)) == '{') {
if ((c = collect(c, '}')) != '\0')
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
goto pack;
}
break;
case '`':
case '\'':
case '"':
if ((c = collect(c, c)) != '\0')
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
goto pack;
case '|':
case '&':
case ';':
startl = 1;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* If more chars process them, else return NULL char */
if ((c1 = dual(c)) != '\0')
return (c1);
else
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case '^':
startl = 1;
Run msh through indent
2004-08-04 19:16:54 +00:00
return ('|');
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case '>':
case '<':
diag(c);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case '\n':
nlseen++;
gethere();
startl = 1;
if (multiline || cf & CONTIN) {
if (interactive && e.iop <= iostack) {
Major rework of the directory structure and the entire build system. -Erik
2001-10-24 05:00:29 +00:00
#ifdef CONFIG_FEATURE_COMMAND_EDITING
Run msh through indent
2004-08-04 19:16:54 +00:00
current_prompt = cprompt->value;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#else
Run msh through indent
2004-08-04 19:16:54 +00:00
prs(cprompt->value);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#endif
}
if (cf & CONTIN)
goto loop;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case '(':
case ')':
startl = 1;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
unget(c);
Run msh through indent
2004-08-04 19:16:54 +00:00
pack:
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
while ((c = my_getc(0)) != 0 && !any(c, "`$ '\"\t;&<>()|^\n")) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (e.linep >= elinep)
err("word too long");
else
*e.linep++ = c;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
};
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
unget(c);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
if (any(c, "\"'`$"))
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
goto loop;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
*e.linep++ = '\0';
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
if (atstart && (c = rlookup(line)) != 0) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
startl = 1;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
yylval.cp = strsave(line, areanum);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (WORD);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static int collect(c, c1)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c, c1;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
char s[2];
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF8(("COLLECT: enter, c=%d, c1=%d\n", c, c1));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
*e.linep++ = c;
while ((c = my_getc(c1)) != c1) {
if (c == 0) {
unget(c);
s[0] = c1;
s[1] = 0;
Run msh through indent
2004-08-04 19:16:54 +00:00
prs("no closing ");
yyerror(s);
return (YYERRCODE);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
if (interactive && c == '\n' && e.iop <= iostack) {
Major rework of the directory structure and the entire build system. -Erik
2001-10-24 05:00:29 +00:00
#ifdef CONFIG_FEATURE_COMMAND_EDITING
Run msh through indent
2004-08-04 19:16:54 +00:00
current_prompt = cprompt->value;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#else
Run msh through indent
2004-08-04 19:16:54 +00:00
prs(cprompt->value);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#endif
}
*e.linep++ = c;
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
*e.linep++ = c;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF8(("COLLECT: return 0, line is %s\n", line));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* "multiline commands" helper func */
/* see if next 2 chars form a shell multiline */
Run msh through indent
2004-08-04 19:16:54 +00:00
static int dual(c)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
char s[3];
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp = s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF8(("DUAL: enter, c=%d\n", c));
*cp++ = c; /* c is the given "peek" char */
*cp++ = my_getc(0); /* get next char of input */
*cp = 0; /* add EOS marker */
c = rlookup(s); /* see if 2 chars form a shell multiline */
if (c == 0)
unget(*--cp); /* String is not a shell multiline, put peek char back */
return (c); /* String is multiline, return numeric multiline (restab) code */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void diag(ec)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int ec;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
DBGPRINTF8(("DIAG: enter, ec=%d\n", ec));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
c = my_getc(0);
if (c == '>' || c == '<') {
if (c != ec)
zzerr();
Run msh through indent
2004-08-04 19:16:54 +00:00
yylval.i = ec == '>' ? IOWRITE | IOCAT : IOHERE;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
c = my_getc(0);
} else
Run msh through indent
2004-08-04 19:16:54 +00:00
yylval.i = ec == '>' ? IOWRITE : IOREAD;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (c != '&' || yylval.i == IOHERE)
unget(c);
else
yylval.i |= IODUP;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *tree(size)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
unsigned size;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((t = getcell(size)) == NULL) {
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF2(("TREE: getcell(%d) failed!\n", size));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
prs("command line too complicated\n");
fail();
/* NOTREACHED */
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (t);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/* VARARGS1 */
/* ARGSUSED */
/* -------- exec.c -------- */
/*
* execute tree
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int execute(t, pin, pout, act)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int *pin, *pout;
int act;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t1;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
volatile int i, rv, a;
char *cp, **wp, **wp2;
struct var *vp;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
struct op *outtree_save;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct brkcon bc;
#if __GNUC__
/* Avoid longjmp clobbering */
(void) &wp;
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
#endif
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if (t == NULL) {
DBGPRINTF4(("EXECUTE: enter, t==null, returning.\n"));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
DBGPRINTF(("EXECUTE: t=0x%x, t->type=%d (%s), t->words is %s\n", t,
t->type, T_CMD_NAMES[t->type],
((t->words == NULL) ? "NULL" : t->words[0])));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
rv = 0;
a = areanum++;
wp = (wp2 = t->words) != NULL
Run msh through indent
2004-08-04 19:16:54 +00:00
? eval(wp2, t->type == TCOM ? DOALL : DOALL & ~DOKEY)
: NULL;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* Hard to know how many words there are, be careful of garbage pointer values */
/* They are likely to cause "PCI bus fault" errors */
#if 0
DBGPRINTF(("EXECUTE: t->left=0x%x, t->right=0x%x, t->words[1] is %s\n",
t->left, t->right,
((t->words[1] == NULL) ? "NULL" : t->words[1])));
DBGPRINTF7(("EXECUTE: t->words[2] is %s, t->words[3] is %s\n",
((t->words[2] == NULL) ? "NULL" : t->words[2]),
((t->words[3] == NULL) ? "NULL" : t->words[3])));
#endif
Run msh through indent
2004-08-04 19:16:54 +00:00
switch (t->type) {
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
case TDOT:
DBGPRINTF3(("EXECUTE: TDOT\n"));
outtree_save = outtree;
newfile(evalstr(t->words[0], DOALL));
t->left = dowholefile(TLIST, 0);
t->right = NULL;
outtree = outtree_save;
if (t->left)
rv = execute(t->left, pin, pout, 0);
if (t->right)
rv = execute(t->right, pin, pout, 0);
break;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case TPAREN:
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
rv = execute(t->left, pin, pout, 0);
break;
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case TCOM:
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
rv = forkexec(t, pin, pout, act, wp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
break;
case TPIPE:
{
Run msh through indent
2004-08-04 19:16:54 +00:00
int pv[2];
if ((rv = openpipe(pv)) < 0)
break;
pv[0] = remap(pv[0]);
pv[1] = remap(pv[1]);
(void) execute(t->left, pin, pv, 0);
rv = execute(t->right, pv, pout, 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
break;
case TLIST:
(void) execute(t->left, pin, pout, 0);
rv = execute(t->right, pin, pout, 0);
break;
case TASYNC:
Run msh through indent
2004-08-04 19:16:54 +00:00
{
int hinteractive = interactive;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("EXECUTE: TASYNC clause, calling vfork()...\n"));
Run msh through indent
2004-08-04 19:16:54 +00:00
i = vfork();
if (i != 0) {
interactive = hinteractive;
if (i != -1) {
setval(lookup("!"), putn(i));
if (pin != NULL)
closepipe(pin);
if (interactive) {
prs(putn(i));
prs("\n");
}
} else
rv = -1;
setstatus(rv);
} else {
signal(SIGINT, SIG_IGN);
signal(SIGQUIT, SIG_IGN);
if (interactive)
signal(SIGTERM, SIG_DFL);
interactive = 0;
if (pin == NULL) {
close(0);
usage bb_dev_null
2005-10-12 15:34:25 +00:00
open(bb_dev_null, 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
_exit(execute(t->left, pin, pout, FEXEC));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
}
break;
case TOR:
case TAND:
rv = execute(t->left, pin, pout, 0);
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((t1 = t->right) != NULL && (rv == 0) == (t->type == TAND))
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
rv = execute(t1, pin, pout, 0);
break;
case TFOR:
if (wp == NULL) {
Run msh through indent
2004-08-04 19:16:54 +00:00
wp = dolv + 1;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((i = dolc) < 0)
i = 0;
} else {
i = -1;
Run msh through indent
2004-08-04 19:16:54 +00:00
while (*wp++ != NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
vp = lookup(t->str);
while (setjmp(bc.brkpt))
if (isbreak)
goto broken;
brkset(&bc);
for (t1 = t->left; i-- && *wp != NULL;) {
setval(vp, *wp++);
rv = execute(t1, pin, pout, 0);
}
brklist = brklist->nextlev;
break;
case TWHILE:
case TUNTIL:
while (setjmp(bc.brkpt))
if (isbreak)
goto broken;
brkset(&bc);
t1 = t->left;
while ((execute(t1, pin, pout, 0) == 0) == (t->type == TWHILE))
rv = execute(t->right, pin, pout, 0);
brklist = brklist->nextlev;
break;
case TIF:
case TELIF:
Run msh through indent
2004-08-04 19:16:54 +00:00
if (t->right != NULL) {
rv = !execute(t->left, pin, pout, 0) ?
execute(t->right->left, pin, pout, 0) :
execute(t->right->right, pin, pout, 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
break;
case TCASE:
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((cp = evalstr(t->str, DOSUB | DOTRIM)) == 0)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
cp = "";
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("EXECUTE: TCASE, t->str is %s, cp is %s\n",
((t->str == NULL) ? "NULL" : t->str),
((cp == NULL) ? "NULL" : cp)));
if ((t1 = findcase(t->left, cp)) != NULL) {
DBGPRINTF7(("EXECUTE: TCASE, calling execute(t=0x%x, t1=0x%x)...\n", t, t1));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
rv = execute(t1, pin, pout, 0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("EXECUTE: TCASE, back from execute(t=0x%x, t1=0x%x)...\n", t, t1));
}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
break;
case TBRACE:
/*
if (iopp = t->ioact)
while (*iopp)
if (iosetup(*iopp++, pin!=NULL, pout!=NULL)) {
rv = -1;
break;
}
*/
if (rv >= 0 && (t1 = t->left))
rv = execute(t1, pin, pout, 0);
break;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
};
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
broken:
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t->words = wp2;
isbreak = 0;
freehere(areanum);
freearea(areanum);
areanum = a;
if (interactive && intr) {
closeall();
fail();
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((i = trapset) != 0) {
trapset = 0;
runtrap(i);
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("EXECUTE: returning from t=0x%x, rv=%d\n", t, rv));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (rv);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
static int
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
forkexec(REGISTER struct op *t, int *pin, int *pout, int act, char **wp)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
pid_t newpid;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int i, rv;
Run msh through indent
2004-08-04 19:16:54 +00:00
int (*shcom) (struct op *) = NULL;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int f;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *cp = NULL;
struct ioword **iopp;
int resetsig;
char **owp;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
int forked = 0;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int *hpin = pin;
int *hpout = pout;
char *hwp;
int hinteractive;
int hintr;
Run msh through indent
2004-08-04 19:16:54 +00:00
struct brkcon *hbrklist;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int hexecflg;
#if __GNUC__
/* Avoid longjmp clobbering */
(void) &pin;
(void) &pout;
(void) &wp;
(void) &shcom;
(void) &cp;
(void) &resetsig;
(void) &owp;
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
#endif
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("FORKEXEC: t=0x%x, pin 0x%x, pout 0x%x, act %d\n", t, pin,
pout, act));
DBGPRINTF7(("FORKEXEC: t->words is %s\n",
((t->words == NULL) ? "NULL" : t->words[0])));
/* Hard to know how many words there are, be careful of garbage pointer values */
/* They are likely to cause "PCI bus fault" errors */
#if 0
DBGPRINTF7(("FORKEXEC: t->words is %s, t->words[1] is %s\n",
((t->words == NULL) ? "NULL" : t->words[0]),
((t->words == NULL) ? "NULL" : t->words[1])));
DBGPRINTF7(("FORKEXEC: wp is %s, wp[1] is %s\n",
((wp == NULL) ? "NULL" : wp[0]),
((wp[1] == NULL) ? "NULL" : wp[1])));
DBGPRINTF7(("FORKEXEC: wp2 is %s, wp[3] is %s\n",
((wp[2] == NULL) ? "NULL" : wp[2]),
((wp[3] == NULL) ? "NULL" : wp[3])));
#endif
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
owp = wp;
resetsig = 0;
Run msh through indent
2004-08-04 19:16:54 +00:00
rv = -1; /* system-detected error */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (t->type == TCOM) {
Run msh through indent
2004-08-04 19:16:54 +00:00
while ((cp = *wp++) != NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
cp = *wp;
/* strip all initial assignments */
/* not correct wrt PATH=yyy command etc */
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if (flag['x']) {
DBGPRINTF9(("FORKEXEC: echo'ing, cp=0x%x, wp=0x%x, owp=0x%x\n",
cp, wp, owp));
Run msh through indent
2004-08-04 19:16:54 +00:00
echo(cp ? wp : owp);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
#if 0
DBGPRINTF9(("FORKEXEC: t->words is %s, t->words[1] is %s\n",
((t->words == NULL) ? "NULL" : t->words[0]),
((t->words == NULL) ? "NULL" : t->words[1])));
DBGPRINTF9(("FORKEXEC: wp is %s, wp[1] is %s\n",
((wp == NULL) ? "NULL" : wp[0]),
((wp == NULL) ? "NULL" : wp[1])));
#endif
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (cp == NULL && t->ioact == NULL) {
Run msh through indent
2004-08-04 19:16:54 +00:00
while ((cp = *owp++) != NULL && assign(cp, COPYV));
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("FORKEXEC: returning setstatus()\n"));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (setstatus(0));
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
} else if (cp != NULL) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
shcom = inbuilt(cp);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
t->words = wp;
f = act;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
#if 0
DBGPRINTF3(("FORKEXEC: t->words is %s, t->words[1] is %s\n",
((t->words == NULL) ? "NULL" : t->words[0]),
((t->words == NULL) ? "NULL" : t->words[1])));
#endif
DBGPRINTF(("FORKEXEC: shcom 0x%x, f&FEXEC 0x%x, owp 0x%x\n", shcom,
f & FEXEC, owp));
if (shcom == NULL && (f & FEXEC) == 0) {
/* Save values in case the child process alters them */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
hpin = pin;
hpout = pout;
hwp = *wp;
hinteractive = interactive;
hintr = intr;
hbrklist = brklist;
hexecflg = execflg;
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF3(("FORKEXEC: calling vfork()...\n"));
newpid = vfork();
if (newpid == -1) {
DBGPRINTF(("FORKEXEC: ERROR, unable to vfork()!\n"));
return (-1);
}
if (newpid > 0) { /* Parent */
/* Restore values */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
pin = hpin;
pout = hpout;
*wp = hwp;
interactive = hinteractive;
intr = hintr;
brklist = hbrklist;
execflg = hexecflg;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* moved up
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (i == -1)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
return(rv);
*/
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (pin != NULL)
closepipe(pin);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
return (pout == NULL ? setstatus(waitfor(newpid, 0)) : 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* Must be the child process, pid should be 0 */
DBGPRINTF(("FORKEXEC: child process, shcom=0x%x\n", shcom));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (interactive) {
signal(SIGINT, SIG_IGN);
signal(SIGQUIT, SIG_IGN);
resetsig = 1;
}
interactive = 0;
intr = 0;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
forked = 1;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
brklist = 0;
execflg = 0;
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (owp != NULL)
while ((cp = *owp++) != NULL && assign(cp, COPYV))
if (shcom == NULL)
export(lookup(cp));
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#ifdef COMPIPE
if ((pin != NULL || pout != NULL) && shcom != NULL && shcom != doexec) {
err("piping to/from shell builtins not yet done");
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if (forked)
_exit(-1);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (-1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
#endif
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (pin != NULL) {
dup2(pin[0], 0);
closepipe(pin);
}
if (pout != NULL) {
dup2(pout[1], 1);
closepipe(pout);
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((iopp = t->ioact) != NULL) {
if (shcom != NULL && shcom != doexec) {
prs(cp);
err(": cannot redirect shell command");
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if (forked)
_exit(-1);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (-1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
while (*iopp)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if (iosetup(*iopp++, pin != NULL, pout != NULL)) {
if (forked)
_exit(rv);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (rv);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
}
if (shcom) {
i = setstatus((*shcom) (t));
if (forked)
_exit(i);
DBGPRINTF(("FORKEXEC: returning i=%d\n", i));
return (i);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* should use FIOCEXCL */
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = FDBASE; i < NOFILE; i++)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
close(i);
if (resetsig) {
signal(SIGINT, SIG_DFL);
signal(SIGQUIT, SIG_DFL);
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (t->type == TPAREN)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
_exit(execute(t->left, NOPIPE, NOPIPE, FEXEC));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (wp[0] == NULL)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
_exit(0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Jonas Holmberg from axis dot com writes: This patch makes msh handle variable expansion within backticks more correctly. Current behaviour (wrong): -------------------------- BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` hello $ New behaviour (correct): ------------------------ BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` `echo hello` $ The current behaviour (wrong according to standards) was actually my fault. msh handles backticks by executing a subshell (which makes it work on MMU-less systems). Executing a subshell makes it hard to only expand variables once in the parent. Therefore I export all variables that will be expanded within the backticks and let the subshell handle the expansion instead. The bug was found while searching for security leaks in CGI-scripts. Current behaviour of msh makes it easy to expand backticks by mistake in $QUERY_STRING. I recommend appling the patch before release of bb 1.00. /Jonas
2004-09-02 23:13:10 +00:00
cp = rexecve(wp[0], wp, makenv(0, NULL));
Run msh through indent
2004-08-04 19:16:54 +00:00
prs(wp[0]);
prs(": ");
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
err(cp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (!execflg)
trap[0] = NULL;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("FORKEXEC: calling leave(), pid=%d\n", newpid));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
leave();
/* NOTREACHED */
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
_exit(1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* 0< 1> are ignored as required
* within pipelines.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int iosetup(iop, pipein, pipeout)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioword *iop;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int pipein, pipeout;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int u = -1;
Run msh through indent
2004-08-04 19:16:54 +00:00
char *cp = NULL, *msg;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("IOSETUP: iop 0x%x, pipein 0x%x, pipeout 0x%x\n", iop,
pipein, pipeout));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (iop->io_unit == IODEFAULT) /* take default */
Run msh through indent
2004-08-04 19:16:54 +00:00
iop->io_unit = iop->io_flag & (IOREAD | IOHERE) ? 0 : 1;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (pipein && iop->io_unit == 0)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (pipeout && iop->io_unit == 1)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
msg = iop->io_flag & (IOREAD | IOHERE) ? "open" : "create";
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((iop->io_flag & IOHERE) == 0) {
cp = iop->io_name;
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((cp = evalstr(cp, DOSUB | DOTRIM)) == NULL)
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (iop->io_flag & IODUP) {
if (cp[1] || (!isdigit(*cp) && *cp != '-')) {
prs(cp);
err(": illegal >& argument");
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
if (*cp == '-')
iop->io_flag = IOCLOSE;
Run msh through indent
2004-08-04 19:16:54 +00:00
iop->io_flag &= ~(IOREAD | IOWRITE);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
switch (iop->io_flag) {
case IOREAD:
u = open(cp, 0);
break;
case IOHERE:
Run msh through indent
2004-08-04 19:16:54 +00:00
case IOHERE | IOXHERE:
u = herein(iop->io_name, iop->io_flag & IOXHERE);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
cp = "here file";
break;
Run msh through indent
2004-08-04 19:16:54 +00:00
case IOWRITE | IOCAT:
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((u = open(cp, 1)) >= 0) {
Run msh through indent
2004-08-04 19:16:54 +00:00
lseek(u, (long) 0, 2);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
break;
}
case IOWRITE:
u = creat(cp, 0666);
break;
case IODUP:
Run msh through indent
2004-08-04 19:16:54 +00:00
u = dup2(*cp - '0', iop->io_unit);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
break;
case IOCLOSE:
close(iop->io_unit);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
if (u < 0) {
prs(cp);
prs(": cannot ");
warn(msg);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
} else {
if (u != iop->io_unit) {
dup2(u, iop->io_unit);
close(u);
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void echo(wp)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char **wp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
prs("+");
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = 0; wp[i]; i++) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (i)
prs(" ");
prs(wp[i]);
}
prs("\n");
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op **find1case(t, w)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op *t;
char *w;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t1;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op **tp;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char **wp, *cp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if (t == NULL) {
DBGPRINTF3(("FIND1CASE: enter, t==NULL, returning.\n"));
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((struct op **) NULL);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
DBGPRINTF3(("FIND1CASE: enter, t->type=%d (%s)\n", t->type,
T_CMD_NAMES[t->type]));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (t->type == TLIST) {
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if ((tp = find1case(t->left, w)) != NULL) {
DBGPRINTF3(("FIND1CASE: found one to the left, returning tp=0x%x\n", tp));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (tp);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
t1 = t->right; /* TPAT */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
} else
t1 = t;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (wp = t1->words; *wp;)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if ((cp = evalstr(*wp++, DOSUB)) && gmatch(w, cp)) {
DBGPRINTF3(("FIND1CASE: returning &t1->left= 0x%x.\n",
&t1->left));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (&t1->left);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
DBGPRINTF(("FIND1CASE: returning NULL\n"));
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((struct op **) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct op *findcase(t, w)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op *t;
char *w;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op **tp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((tp = find1case(t, w)) != NULL ? *tp : (struct op *) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Enter a new loop level (marked for break/continue).
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static void brkset(bc)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct brkcon *bc;
{
bc->nextlev = brklist;
brklist = bc;
}
/*
* Wait for the last process created.
* Print a message for each process found
* that was killed by a signal.
* Ignore interrupt signals while waiting
* unless `canintr' is true.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int waitfor(lastpid, canintr)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int lastpid;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int canintr;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int pid, rv;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int s;
int oheedint = heedint;
heedint = 0;
rv = 0;
do {
pid = wait(&s);
if (pid == -1) {
if (errno != EINTR || canintr)
break;
} else {
if ((rv = WAITSIG(s)) != 0) {
if (rv < NSIGNAL) {
if (signame[rv] != NULL) {
if (pid != lastpid) {
prn(pid);
prs(": ");
}
prs(signame[rv]);
}
} else {
if (pid != lastpid) {
prn(pid);
prs(": ");
}
Run msh through indent
2004-08-04 19:16:54 +00:00
prs("Signal ");
prn(rv);
prs(" ");
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
if (WAITCORE(s))
prs(" - core dumped");
if (rv >= NSIGNAL || signame[rv])
prs("\n");
rv = -1;
} else
rv = WAITVAL(s);
}
} while (pid != lastpid);
heedint = oheedint;
if (intr) {
if (interactive) {
if (canintr)
intr = 0;
} else {
Run msh through indent
2004-08-04 19:16:54 +00:00
if (exstat == 0)
exstat = rv;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
onintr(0);
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (rv);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int setstatus(s)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
exstat = s;
setval(lookup("?"), putn(s));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (s);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* PATH-searching interface to execve.
* If getenv("PATH") were kept up-to-date,
* execvp might be used.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *rexecve(c, v, envp)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *c, **v, **envp;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
REGISTER char *sp, *tp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int eacces = 0, asis = 0;
Major rework of the directory structure and the entire build system. -Erik
2001-10-24 05:00:29 +00:00
#ifdef CONFIG_FEATURE_SH_STANDALONE_SHELL
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
char *name = c;
Run msh through indent
2004-08-04 19:16:54 +00:00
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
optind = 1;
if (find_applet_by_name(name)) {
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
/* We have to exec here since we vforked. Running
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
* run_applet_by_name() won't work and bad things
* will happen. */
execve("/proc/self/exe", v, envp);
execve("busybox", v, envp);
}
#endif
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("REXECVE: c=0x%x, v=0x%x, envp=0x%x\n", c, v, envp));
Run msh through indent
2004-08-04 19:16:54 +00:00
sp = any('/', c) ? "" : path->value;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
asis = *sp == '\0';
while (asis || *sp != '\0') {
asis = 0;
tp = e.linep;
for (; *sp != '\0'; tp++)
if ((*tp = *sp++) == ':') {
asis = *sp == '\0';
break;
}
if (tp != e.linep)
*tp++ = '/';
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = 0; (*tp++ = c[i++]) != '\0';);
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF3(("REXECVE: e.linep is %s\n", e.linep));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
execve(e.linep, v, envp);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
switch (errno) {
case ENOEXEC:
*v = e.linep;
tp = *--v;
*v = e.linep;
move all "-/bin/sh" "/bin/sh" and "sh" to libbb/messages.c file as one constant. Vodz last_patch_107
2003-09-02 02:36:18 +00:00
execve(DEFAULT_SHELL, v, envp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
*v = tp;
Run msh through indent
2004-08-04 19:16:54 +00:00
return ("no Shell");
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case ENOMEM:
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((char *) bb_msg_memory_exhausted);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case E2BIG:
Run msh through indent
2004-08-04 19:16:54 +00:00
return ("argument list too long");
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
case EACCES:
eacces++;
break;
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (errno == ENOENT ? "not found" : "cannot execute");
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Run the command produced by generator `f'
* applied to stream `arg'.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int run(struct ioarg *argp, int (*f) (struct ioarg *))
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
struct op *otree;
struct wdblock *swdlist;
struct wdblock *siolist;
jmp_buf ev, rt;
xint *ofail;
int rv;
#if __GNUC__
/* Avoid longjmp clobbering */
(void) &rv;
#endif
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("RUN: enter, areanum %d, outtree 0x%x, failpt 0x%x\n",
areanum, outtree, failpt));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
areanum++;
swdlist = wdlist;
siolist = iolist;
otree = outtree;
ofail = failpt;
rv = -1;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (newenv(setjmp(errpt = ev)) == 0) {
wdlist = 0;
iolist = 0;
pushio(argp, f);
e.iobase = e.iop;
yynerrs = 0;
if (setjmp(failpt = rt) == 0 && yyparse() == 0)
rv = execute(outtree, NOPIPE, NOPIPE, 0);
quitenv();
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
} else {
DBGPRINTF(("RUN: error from newenv()!\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
wdlist = swdlist;
iolist = siolist;
failpt = ofail;
outtree = otree;
freearea(areanum--);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
return (rv);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/* -------- do.c -------- */
/*
* built-in commands: doX
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int dohelp(struct op *t)
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
{
int col;
const struct builtincmd *x;
printf("\nBuilt-in commands:\n");
printf("-------------------\n");
Run msh through indent
2004-08-04 19:16:54 +00:00
for (col = 0, x = builtincmds; x->builtinfunc != NULL; x++) {
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
if (!x->name)
continue;
col += printf("%s%s", ((col == 0) ? "\t" : " "), x->name);
if (col > 60) {
printf("\n");
col = 0;
}
}
Major rework of the directory structure and the entire build system. -Erik
2001-10-24 05:00:29 +00:00
#ifdef CONFIG_FEATURE_SH_STANDALONE_SHELL
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
{
int i;
const struct BB_applet *applet;
extern const struct BB_applet applets[];
extern const size_t NUM_APPLETS;
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = 0, applet = applets; i < NUM_APPLETS; applet++, i++) {
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
if (!applet->name)
continue;
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
col += printf("%s%s", ((col == 0) ? "\t" : " "), applet->name);
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
if (col > 60) {
printf("\n");
col = 0;
}
}
}
#endif
printf("\n\n");
return EXIT_SUCCESS;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int dolabel(struct op *t)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int dochdir(t)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp, *er;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((cp = t->words[1]) == NULL && (cp = homedir->value) == NULL)
er = ": no home directory";
Run msh through indent
2004-08-04 19:16:54 +00:00
else if (chdir(cp) < 0)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
er = ": bad directory";
else
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
prs(cp != NULL ? cp : "cd");
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
err(er);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int doshift(t)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int n;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
n = t->words[1] ? getn(t->words[1]) : 1;
if (dolc < n) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
err("nothing to shift");
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
dolv[n] = dolv[0];
dolv += n;
dolc -= n;
setval(lookup("#"), putn(dolc));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* execute login and newgrp directly
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int dologin(t)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op *t;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (interactive) {
signal(SIGINT, SIG_DFL);
signal(SIGQUIT, SIG_DFL);
}
Jonas Holmberg from axis dot com writes: This patch makes msh handle variable expansion within backticks more correctly. Current behaviour (wrong): -------------------------- BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` hello $ New behaviour (correct): ------------------------ BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` `echo hello` $ The current behaviour (wrong according to standards) was actually my fault. msh handles backticks by executing a subshell (which makes it work on MMU-less systems). Executing a subshell makes it hard to only expand variables once in the parent. Therefore I export all variables that will be expanded within the backticks and let the subshell handle the expansion instead. The bug was found while searching for security leaks in CGI-scripts. Current behaviour of msh makes it easy to expand backticks by mistake in $QUERY_STRING. I recommend appling the patch before release of bb 1.00. /Jonas
2004-09-02 23:13:10 +00:00
cp = rexecve(t->words[0], t->words, makenv(0, NULL));
Run msh through indent
2004-08-04 19:16:54 +00:00
prs(t->words[0]);
prs(": ");
err(cp);
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int doumask(t)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i, n;
REGISTER char *cp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((cp = t->words[1]) == NULL) {
i = umask(0);
umask(i);
Run msh through indent
2004-08-04 19:16:54 +00:00
for (n = 3 * 4; (n -= 3) >= 0;)
putc('0' + ((i >> n) & 07), stderr);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
putc('\n', stderr);
} else {
Run msh through indent
2004-08-04 19:16:54 +00:00
for (n = 0; *cp >= '0' && *cp <= '9'; cp++)
n = n * 8 + (*cp - '0');
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
umask(n);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int doexec(t)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
jmp_buf ex;
xint *ofail;
t->ioact = NULL;
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = 0; (t->words[i] = t->words[i + 1]) != NULL; i++);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (i == 0)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
execflg = 1;
ofail = failpt;
if (setjmp(failpt = ex) == 0)
execute(t, NOPIPE, NOPIPE, FEXEC);
failpt = ofail;
execflg = 0;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int dodot(t)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op *t;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
REGISTER char *sp, *tp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *cp;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
int maltmp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("DODOT: enter, t=0x%x, tleft 0x%x, tright 0x%x, e.linep is %s\n", t, t->left, t->right, ((e.linep == NULL) ? "NULL" : e.linep)));
if ((cp = t->words[1]) == NULL) {
DBGPRINTF(("DODOT: bad args, ret 0\n"));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
} else {
DBGPRINTF(("DODOT: cp is %s\n", cp));
}
Run msh through indent
2004-08-04 19:16:54 +00:00
sp = any('/', cp) ? ":" : path->value;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("DODOT: sp is %s, e.linep is %s\n",
((sp == NULL) ? "NULL" : sp),
((e.linep == NULL) ? "NULL" : e.linep)));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
while (*sp) {
tp = e.linep;
while (*sp && (*tp = *sp++) != ':')
tp++;
if (tp != e.linep)
*tp++ = '/';
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = 0; (*tp++ = cp[i++]) != '\0';);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* Original code */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((i = open(e.linep, 0)) >= 0) {
exstat = 0;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
maltmp = remap(i);
DBGPRINTF(("DODOT: remap=%d, exstat=%d, e.iofd %d, i %d, e.linep is %s\n", maltmp, exstat, e.iofd, i, e.linep));
next(maltmp); /* Basically a PUSHIO */
DBGPRINTF(("DODOT: returning exstat=%d\n", exstat));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (exstat);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
} /* While */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
prs(cp);
err(": not found");
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
return (-1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int dowait(t)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op *t;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
REGISTER char *cp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((cp = t->words[1]) != NULL) {
i = getn(cp);
if (i == 0)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
} else
i = -1;
setstatus(waitfor(i, 1));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int doread(t)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op *t;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp, **wp;
REGISTER int nb = 0;
REGISTER int nl = 0;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (t->words[1] == NULL) {
err("Usage: read name ...");
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
for (wp = t->words + 1; *wp; wp++) {
for (cp = e.linep; !nl && cp < elinep - 1; cp++)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((nb = read(0, cp, sizeof(*cp))) != sizeof(*cp) ||
Run msh through indent
2004-08-04 19:16:54 +00:00
(nl = (*cp == '\n')) || (wp[1] && any(*cp, ifs->value)))
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
break;
*cp = 0;
if (nb <= 0)
break;
setval(lookup(*wp), e.linep);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (nb <= 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int doeval(t)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Run msh through indent
2004-08-04 19:16:54 +00:00
return (RUN(awordlist, t->words + 1, wdchar));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int dotrap(t)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int n, i;
REGISTER int resetsig;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (t->words[1] == NULL) {
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = 0; i <= _NSIG; i++)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (trap[i]) {
prn(i);
prs(": ");
prs(trap[i]);
prs("\n");
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
resetsig = isdigit(*t->words[1]);
for (i = resetsig ? 1 : 2; t->words[i] != NULL; ++i) {
n = getsig(t->words[i]);
freecell(trap[n]);
trap[n] = 0;
if (!resetsig) {
if (*t->words[1] != '\0') {
trap[n] = strsave(t->words[1], 0);
setsig(n, sig);
} else
setsig(n, SIG_IGN);
} else {
if (interactive)
if (n == SIGINT)
setsig(n, onintr);
else
Run msh through indent
2004-08-04 19:16:54 +00:00
setsig(n, n == SIGQUIT ? SIG_IGN : SIG_DFL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
else
setsig(n, SIG_DFL);
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int getsig(s)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *s;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int n;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((n = getn(s)) < 0 || n > _NSIG) {
err("trap: bad signal number");
n = 0;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (n);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
static void setsig(REGISTER int n, sighandler_t f)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
if (n == 0)
return;
if (signal(n, SIG_IGN) != SIG_IGN || ourtrap[n]) {
ourtrap[n] = 1;
signal(n, f);
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int getn(as)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *as;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s;
REGISTER int n, m;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
s = as;
m = 1;
if (*s == '-') {
m = -1;
s++;
}
for (n = 0; isdigit(*s); s++)
Run msh through indent
2004-08-04 19:16:54 +00:00
n = (n * 10) + (*s - '0');
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (*s) {
prs(as);
err(": bad number");
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (n * m);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int dobreak(t)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op *t;
{
Run msh through indent
2004-08-04 19:16:54 +00:00
return (brkcontin(t->words[1], 1));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int docontinue(t)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op *t;
{
Run msh through indent
2004-08-04 19:16:54 +00:00
return (brkcontin(t->words[1], 0));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int brkcontin(cp, val)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int val;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct brkcon *bc;
REGISTER int nl;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
nl = cp == NULL ? 1 : getn(cp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (nl <= 0)
nl = 999;
do {
if ((bc = brklist) == NULL)
break;
brklist = bc->nextlev;
} while (--nl);
if (nl) {
err("bad break/continue level");
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
isbreak = val;
longjmp(bc->brkpt, 1);
/* NOTREACHED */
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int doexit(t)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op *t;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
execflg = 0;
if ((cp = t->words[1]) != NULL)
setstatus(getn(cp));
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("DOEXIT: calling leave(), t=0x%x\n", t));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
leave();
/* NOTREACHED */
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int doexport(t)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op *t;
{
Run msh through indent
2004-08-04 19:16:54 +00:00
rdexp(t->words + 1, export, EXPORT);
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int doreadonly(t)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct op *t;
{
Run msh through indent
2004-08-04 19:16:54 +00:00
rdexp(t->words + 1, ronly, RONLY);
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void rdexp(char **wp, void (*f) (struct var *), int key)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF6(("RDEXP: enter, wp=0x%x, func=0x%x, key=%d\n", wp, f, key));
DBGPRINTF6(("RDEXP: *wp=%s\n", *wp));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (*wp != NULL) {
Allow assignments in the export and readonly commands.
2001-07-12 19:39:59 +00:00
for (; *wp != NULL; wp++) {
if (isassign(*wp)) {
char *cp;
Run msh through indent
2004-08-04 19:16:54 +00:00
Allow assignments in the export and readonly commands.
2001-07-12 19:39:59 +00:00
assign(*wp, COPYV);
Run msh through indent
2004-08-04 19:16:54 +00:00
for (cp = *wp; *cp != '='; cp++);
Allow assignments in the export and readonly commands.
2001-07-12 19:39:59 +00:00
*cp = '\0';
}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (checkname(*wp))
Run msh through indent
2004-08-04 19:16:54 +00:00
(*f) (lookup(*wp));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
else
badid(*wp);
Allow assignments in the export and readonly commands.
2001-07-12 19:39:59 +00:00
}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
} else
putvlist(key, 1);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void badid(s)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
prs(s);
err(": bad identifier");
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int doset(t)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct op *t;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct var *vp;
REGISTER char *cp;
REGISTER int n;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((cp = t->words[1]) == NULL) {
for (vp = vlist; vp; vp = vp->next)
varput(vp->name, 1);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
if (*cp == '-') {
/* bad: t->words++; */
Run msh through indent
2004-08-04 19:16:54 +00:00
for (n = 0; (t->words[n] = t->words[n + 1]) != NULL; n++);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (*++cp == 0)
flag['x'] = flag['v'] = 0;
else
for (; *cp; cp++)
switch (*cp) {
case 'e':
if (!interactive)
flag['e']++;
break;
default:
Run msh through indent
2004-08-04 19:16:54 +00:00
if (*cp >= 'a' && *cp <= 'z')
flag[(int) *cp]++;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
break;
}
setdash();
}
if (t->words[1]) {
t->words[0] = dolv[0];
Run msh through indent
2004-08-04 19:16:54 +00:00
for (n = 1; t->words[n]; n++)
setarea((char *) t->words[n], 0);
dolc = n - 1;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
dolv = t->words;
setval(lookup("#"), putn(dolc));
Run msh through indent
2004-08-04 19:16:54 +00:00
setarea((char *) (dolv - 1), 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void varput(s, out)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int out;
{
Add support for underscores in variable names.
2001-08-06 14:14:18 +00:00
if (isalnum(*s) || *s == '_') {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
write(out, s, strlen(s));
write(out, "\n", 1);
}
}
/*
* Copyright (c) 1999 Herbert Xu <herbert@debian.org>
* This file contains code for the times builtin.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int dotimes(struct op *t)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
struct tms buf;
long int clk_tck = sysconf(_SC_CLK_TCK);
times(&buf);
printf("%dm%fs %dm%fs\n%dm%fs %dm%fs\n",
Run msh through indent
2004-08-04 19:16:54 +00:00
(int) (buf.tms_utime / clk_tck / 60),
((double) buf.tms_utime) / clk_tck,
(int) (buf.tms_stime / clk_tck / 60),
((double) buf.tms_stime) / clk_tck,
(int) (buf.tms_cutime / clk_tck / 60),
((double) buf.tms_cutime) / clk_tck,
(int) (buf.tms_cstime / clk_tck / 60),
((double) buf.tms_cstime) / clk_tck);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
return 0;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int (*inbuilt(char *s)) (struct op *) {
Add in a shell tagline (per lash/hush behavior) to make it easier to know which shell is in use. Add in 'help' to list available builtins, and fixup msh so it can do STANDALONE_SHELL. -Erik
2001-07-07 00:05:55 +00:00
const struct builtincmd *bp;
for (bp = builtincmds; bp->name != NULL; bp++)
if (strcmp(bp->name, s) == 0)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (bp->builtinfunc);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
return (NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/* -------- eval.c -------- */
/*
* ${}
* `command`
* blank interpretation
* quoting
* glob
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static char **eval(char **ap, int f)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
struct wdblock *wb;
char **wp;
char **wf;
jmp_buf ev;
#if __GNUC__
/* Avoid longjmp clobbering */
(void) &wp;
(void) &ap;
#endif
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF4(("EVAL: enter, f=%d\n", f));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
wp = NULL;
wb = NULL;
wf = NULL;
if (newenv(setjmp(errpt = ev)) == 0) {
while (*ap && isassign(*ap))
expand(*ap++, &wb, f & ~DOGLOB);
if (flag['k']) {
for (wf = ap; *wf; wf++) {
if (isassign(*wf))
expand(*wf, &wb, f & ~DOGLOB);
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
for (wb = addword((char *) 0, wb); *ap; ap++) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (!flag['k'] || !isassign(*ap))
expand(*ap, &wb, f & ~DOKEY);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
wb = addword((char *) 0, wb);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
wp = getwords(wb);
quitenv();
} else
gflg = 1;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
return (gflg ? (char **) NULL : wp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Make the exported environment from the exported
* names in the dictionary. Keyword assignments
* will already have been done.
*/
Jonas Holmberg from axis dot com writes: This patch makes msh handle variable expansion within backticks more correctly. Current behaviour (wrong): -------------------------- BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` hello $ New behaviour (correct): ------------------------ BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` `echo hello` $ The current behaviour (wrong according to standards) was actually my fault. msh handles backticks by executing a subshell (which makes it work on MMU-less systems). Executing a subshell makes it hard to only expand variables once in the parent. Therefore I export all variables that will be expanded within the backticks and let the subshell handle the expansion instead. The bug was found while searching for security leaks in CGI-scripts. Current behaviour of msh makes it easy to expand backticks by mistake in $QUERY_STRING. I recommend appling the patch before release of bb 1.00. /Jonas
2004-09-02 23:13:10 +00:00
static char **makenv(int all, struct wdblock *wb)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct var *vp;
DBGPRINTF5(("MAKENV: enter, all=%d\n", all));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (vp = vlist; vp; vp = vp->next)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if (all || vp->status & EXPORT)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
wb = addword(vp->name, wb);
Run msh through indent
2004-08-04 19:16:54 +00:00
wb = addword((char *) 0, wb);
return (getwords(wb));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *evalstr(cp, f)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *cp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int f;
{
struct wdblock *wb;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF6(("EVALSTR: enter, cp=0x%x, f=%d\n", cp, f));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
wb = NULL;
if (expand(cp, &wb, f)) {
Run msh through indent
2004-08-04 19:16:54 +00:00
if (wb == NULL || wb->w_nword == 0
|| (cp = wb->w_words[0]) == NULL)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
cp = "";
DELETE(wb);
} else
cp = NULL;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (cp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
static int expand(char *cp, REGISTER struct wdblock **wbp, int f)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
jmp_buf ev;
#if __GNUC__
/* Avoid longjmp clobbering */
(void) &cp;
#endif
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF3(("EXPAND: enter, f=%d\n", f));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
gflg = 0;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (cp == NULL)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (!anys("$`'\"", cp) &&
Run msh through indent
2004-08-04 19:16:54 +00:00
!anys(ifs->value, cp) && ((f & DOGLOB) == 0 || !anys("[*?", cp))) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
cp = strsave(cp, areanum);
if (f & DOTRIM)
unquote(cp);
*wbp = addword(cp, *wbp);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
if (newenv(setjmp(errpt = ev)) == 0) {
PUSHIO(aword, cp, strchar);
e.iobase = e.iop;
while ((cp = blank(f)) && gflg == 0) {
e.linep = cp;
cp = strsave(cp, areanum);
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((f & DOGLOB) == 0) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (f & DOTRIM)
unquote(cp);
*wbp = addword(cp, *wbp);
} else
*wbp = glob(cp, *wbp);
}
quitenv();
} else
gflg = 1;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (gflg == 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Blank interpretation and quoting
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *blank(f)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int f;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c, c1;
REGISTER char *sp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int scanequals, foundequals;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF3(("BLANK: enter, f=%d\n", f));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
sp = e.linep;
scanequals = f & DOKEY;
foundequals = 0;
Run msh through indent
2004-08-04 19:16:54 +00:00
loop:
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
switch (c = subgetc('"', foundequals)) {
case 0:
if (sp == e.linep)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
*e.linep++ = 0;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (sp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
default:
if (f & DOBLANK && any(c, ifs->value))
goto loop;
break;
case '"':
case '\'':
scanequals = 0;
if (INSUB())
break;
for (c1 = c; (c = subgetc(c1, 1)) != c1;) {
if (c == 0)
break;
if (c == '\'' || !any(c, "$`\""))
c |= QUOTE;
*e.linep++ = c;
}
c = 0;
}
unget(c);
Add support for underscores in variable names.
2001-08-06 14:14:18 +00:00
if (!isalpha(c) && c != '_')
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
scanequals = 0;
for (;;) {
c = subgetc('"', foundequals);
if (c == 0 ||
Run msh through indent
2004-08-04 19:16:54 +00:00
f & (DOBLANK && any(c, ifs->value)) ||
(!INSUB() && any(c, "\"'"))) {
scanequals = 0;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
unget(c);
if (any(c, "\"'"))
goto loop;
break;
}
if (scanequals) {
if (c == '=') {
foundequals = 1;
Run msh through indent
2004-08-04 19:16:54 +00:00
scanequals = 0;
} else if (!isalnum(c) && c != '_')
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
scanequals = 0;
}
*e.linep++ = c;
}
*e.linep++ = 0;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (sp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Get characters, substituting for ` and $
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int subgetc(ec, quoted)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char ec;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int quoted;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char c;
DBGPRINTF3(("SUBGETC: enter, quoted=%d\n", quoted));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
again:
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
c = my_getc(ec);
if (!INSUB() && ec != '\'') {
if (c == '`') {
if (grave(quoted) == 0)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
e.iop->task = XGRAVE;
goto again;
}
if (c == '$' && (c = dollar(quoted)) == 0) {
e.iop->task = XDOLL;
goto again;
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Prepare to generate the string returned by ${} substitution.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int dollar(quoted)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int quoted;
{
int otask;
struct io *oiop;
char *dolp;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s, c, *cp = NULL;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct var *vp;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF3(("DOLLAR: enter, quoted=%d\n", quoted));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
c = readc();
s = e.linep;
if (c != '{') {
*e.linep++ = c;
Add support for underscores in variable names.
2001-08-06 14:14:18 +00:00
if (isalpha(c) || c == '_') {
Run msh through indent
2004-08-04 19:16:54 +00:00
while ((c = readc()) != 0 && (isalnum(c) || c == '_'))
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (e.linep < elinep)
*e.linep++ = c;
unget(c);
}
c = 0;
} else {
oiop = e.iop;
otask = e.iop->task;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
e.iop->task = XOTHER;
Run msh through indent
2004-08-04 19:16:54 +00:00
while ((c = subgetc('"', 0)) != 0 && c != '}' && c != '\n')
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (e.linep < elinep)
*e.linep++ = c;
if (oiop == e.iop)
e.iop->task = otask;
if (c != '}') {
err("unclosed ${");
gflg++;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
}
if (e.linep >= elinep) {
err("string in ${} too long");
gflg++;
e.linep -= 10;
}
*e.linep = 0;
if (*s)
Run msh through indent
2004-08-04 19:16:54 +00:00
for (cp = s + 1; *cp; cp++)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (any(*cp, "=-+?")) {
c = *cp;
*cp++ = 0;
break;
}
if (s[1] == 0 && (*s == '*' || *s == '@')) {
if (dolc > 1) {
/* currently this does not distinguish $* and $@ */
/* should check dollar */
e.linep = s;
Run msh through indent
2004-08-04 19:16:54 +00:00
PUSHIO(awordlist, dolv + 1, dolchar);
return (0);
} else { /* trap the nasty ${=} */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
s[0] = '1';
s[1] = 0;
}
}
vp = lookup(s);
if ((dolp = vp->value) == null) {
switch (c) {
case '=':
if (isdigit(*s)) {
err("cannot use ${...=...} with $n");
gflg++;
break;
}
setval(vp, cp);
dolp = vp->value;
break;
case '-':
dolp = strsave(cp, areanum);
break;
case '?':
if (*cp == 0) {
prs("missing value for ");
err(s);
} else
err(cp);
gflg++;
break;
}
} else if (c == '+')
dolp = strsave(cp, areanum);
if (flag['u'] && dolp == null) {
prs("unset variable: ");
err(s);
gflg++;
}
e.linep = s;
PUSHIO(aword, dolp, quoted ? qstrchar : strchar);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Run the command in `...` and read its output.
*/
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static int grave(quoted)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int quoted;
{
char *cp;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
int j;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int pf[2];
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
static char child_cmd[LINELIM];
char *src;
char *dest;
int count;
int ignore;
int ignore_once;
char *argument_list[4];
Jonas Holmberg from axis dot com writes: This patch makes msh handle variable expansion within backticks more correctly. Current behaviour (wrong): -------------------------- BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` hello $ New behaviour (correct): ------------------------ BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` `echo hello` $ The current behaviour (wrong according to standards) was actually my fault. msh handles backticks by executing a subshell (which makes it work on MMU-less systems). Executing a subshell makes it hard to only expand variables once in the parent. Therefore I export all variables that will be expanded within the backticks and let the subshell handle the expansion instead. The bug was found while searching for security leaks in CGI-scripts. Current behaviour of msh makes it easy to expand backticks by mistake in $QUERY_STRING. I recommend appling the patch before release of bb 1.00. /Jonas
2004-09-02 23:13:10 +00:00
struct wdblock *wb = NULL;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#if __GNUC__
/* Avoid longjmp clobbering */
(void) &cp;
#endif
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (cp = e.iop->argp->aword; *cp != '`'; cp++)
if (*cp == 0) {
err("no closing `");
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
/* string copy with dollar expansion */
src = e.iop->argp->aword;
dest = child_cmd;
count = 0;
ignore = 0;
ignore_once = 0;
while ((*src != '`') && (count < LINELIM)) {
if (*src == '\'')
ignore = !ignore;
if (*src == '\\')
ignore_once = 1;
if (*src == '$' && !ignore && !ignore_once) {
struct var *vp;
char var_name[LINELIM];
char alt_value[LINELIM];
int var_index = 0;
int alt_index = 0;
char operator = 0;
int braces = 0;
char *value;
src++;
if (*src == '{') {
braces = 1;
src++;
}
var_name[var_index++] = *src++;
applyinf fix for: 0000155: variable expansion with braces in backticks in msh
2005-07-20 18:33:12 +00:00
while (isalnum(*src) || *src=='_')
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
var_name[var_index++] = *src++;
var_name[var_index] = 0;
if (braces) {
switch (*src) {
case '}':
break;
case '-':
case '=':
case '+':
case '?':
Run msh through indent
2004-08-04 19:16:54 +00:00
operator = * src;
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
break;
default:
err("unclosed ${\n");
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
}
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
if (operator) {
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
src++;
while (*src && (*src != '}')) {
alt_value[alt_index++] = *src++;
}
alt_value[alt_index] = 0;
if (*src != '}') {
err("unclosed ${\n");
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
}
}
src++;
}
Jonas Holmberg from axis dot com writes: This patch makes msh handle variable expansion within backticks more correctly. Current behaviour (wrong): -------------------------- BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` hello $ New behaviour (correct): ------------------------ BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` `echo hello` $ The current behaviour (wrong according to standards) was actually my fault. msh handles backticks by executing a subshell (which makes it work on MMU-less systems). Executing a subshell makes it hard to only expand variables once in the parent. Therefore I export all variables that will be expanded within the backticks and let the subshell handle the expansion instead. The bug was found while searching for security leaks in CGI-scripts. Current behaviour of msh makes it easy to expand backticks by mistake in $QUERY_STRING. I recommend appling the patch before release of bb 1.00. /Jonas
2004-09-02 23:13:10 +00:00
if (isalpha(*var_name)) {
/* let subshell handle it instead */
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
Jonas Holmberg from axis dot com writes: This patch makes msh handle variable expansion within backticks more correctly. Current behaviour (wrong): -------------------------- BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` hello $ New behaviour (correct): ------------------------ BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` `echo hello` $ The current behaviour (wrong according to standards) was actually my fault. msh handles backticks by executing a subshell (which makes it work on MMU-less systems). Executing a subshell makes it hard to only expand variables once in the parent. Therefore I export all variables that will be expanded within the backticks and let the subshell handle the expansion instead. The bug was found while searching for security leaks in CGI-scripts. Current behaviour of msh makes it easy to expand backticks by mistake in $QUERY_STRING. I recommend appling the patch before release of bb 1.00. /Jonas
2004-09-02 23:13:10 +00:00
char *namep = var_name;
*dest++ = '$';
if (braces)
*dest++ = '{';
while (*namep)
*dest++ = *namep++;
if (operator) {
char *altp = alt_value;
*dest++ = operator;
while (*altp)
*dest++ = *altp++;
}
if (braces)
*dest++ = '}';
wb = addword(lookup(var_name)->name, wb);
} else {
/* expand */
vp = lookup(var_name);
if (vp->value != null)
value = (operator == '+') ?
alt_value : vp->value;
else if (operator == '?') {
err(alt_value);
return (0);
} else if (alt_index && (operator != '+')) {
value = alt_value;
if (operator == '=')
setval(vp, value);
} else
continue;
while (*value && (count < LINELIM)) {
*dest++ = *value++;
count++;
}
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
}
} else {
*dest++ = *src++;
count++;
ignore_once = 0;
}
}
*dest = '\0';
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (openpipe(pf) < 0)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
while ((i = vfork()) == -1 && errno == EAGAIN);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF3(("GRAVE: i is %d\n", io));
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
if (i < 0) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
closepipe(pf);
Run msh through indent
2004-08-04 19:16:54 +00:00
err((char *) bb_msg_memory_exhausted);
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
if (i != 0) {
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
waitpid(i, NULL, 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
e.iop->argp->aword = ++cp;
close(pf[1]);
Run msh through indent
2004-08-04 19:16:54 +00:00
PUSHIO(afile, remap(pf[0]),
(int (*)(struct ioarg *)) ((quoted) ? qgravechar :
gravechar));
return (1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/* allow trapped signals */
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
/* XXX - Maybe this signal stuff should go as well? */
Run msh through indent
2004-08-04 19:16:54 +00:00
for (j = 0; j <= _NSIG; j++)
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
if (ourtrap[j] && signal(j, SIG_IGN) != SIG_IGN)
signal(j, SIG_DFL);
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
dup2(pf[1], 1);
closepipe(pf);
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
argument_list[0] = (char *) DEFAULT_SHELL;
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
argument_list[1] = "-c";
argument_list[2] = child_cmd;
argument_list[3] = 0;
Jonas Holmberg from axis dot com writes: This patch makes msh handle variable expansion within backticks more correctly. Current behaviour (wrong): -------------------------- BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` hello $ New behaviour (correct): ------------------------ BusyBox v1.00-rc3 (2004.08.26-11:51+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. $ A='`echo hello`' $ echo $A `echo hello` $ echo `echo $A` `echo hello` $ The current behaviour (wrong according to standards) was actually my fault. msh handles backticks by executing a subshell (which makes it work on MMU-less systems). Executing a subshell makes it hard to only expand variables once in the parent. Therefore I export all variables that will be expanded within the backticks and let the subshell handle the expansion instead. The bug was found while searching for security leaks in CGI-scripts. Current behaviour of msh makes it easy to expand backticks by mistake in $QUERY_STRING. I recommend appling the patch before release of bb 1.00. /Jonas
2004-09-02 23:13:10 +00:00
cp = rexecve(argument_list[0], argument_list, makenv(1, wb));
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
prs(argument_list[0]);
prs(": ");
err(cp);
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
_exit(1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *unquote(as)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *as;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((s = as) != NULL)
while (*s)
*s++ &= ~QUOTE;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (as);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/* -------- glob.c -------- */
/*
* glob
*/
#define scopy(x) strsave((x), areanum)
#define BLKSIZ 512
#define NDENT ((BLKSIZ+sizeof(struct dirent)-1)/sizeof(struct dirent))
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct wdblock *cl, *nl;
static char spcl[] = "[?*";
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct wdblock *glob(cp, wb)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *cp;
struct wdblock *wb;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
REGISTER char *pp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (cp == 0)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (wb);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
i = 0;
for (pp = cp; *pp; pp++)
if (any(*pp, spcl))
i++;
else if (!any(*pp & ~QUOTE, spcl))
*pp &= ~QUOTE;
if (i != 0) {
Run msh through indent
2004-08-04 19:16:54 +00:00
for (cl = addword(scopy(cp), (struct wdblock *) 0); anyspcl(cl);
cl = nl) {
nl = newword(cl->w_nword * 2);
for (i = 0; i < cl->w_nword; i++) { /* for each argument */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (pp = cl->w_words[i]; *pp; pp++)
if (any(*pp, spcl)) {
globname(cl->w_words[i], pp);
break;
}
if (*pp == '\0')
nl = addword(scopy(cl->w_words[i]), nl);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = 0; i < cl->w_nword; i++)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
DELETE(cl->w_words[i]);
DELETE(cl);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = 0; i < cl->w_nword; i++)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
unquote(cl->w_words[i]);
Run msh through indent
2004-08-04 19:16:54 +00:00
glob0((char *) cl->w_words, cl->w_nword, sizeof(char *), xstrcmp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (cl->w_nword) {
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = 0; i < cl->w_nword; i++)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
wb = addword(cl->w_words[i], wb);
DELETE(cl);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (wb);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
}
wb = addword(unquote(cp), wb);
Run msh through indent
2004-08-04 19:16:54 +00:00
return (wb);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void globname(we, pp)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *we;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *pp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *np, *cp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *name, *gp, *dp;
int k;
DIR *dirp;
struct dirent *de;
Run msh through indent
2004-08-04 19:16:54 +00:00
char dname[NAME_MAX + 1];
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct stat dbuf;
for (np = we; np != pp; pp--)
if (pp[-1] == '/')
break;
Run msh through indent
2004-08-04 19:16:54 +00:00
for (dp = cp = space((int) (pp - np) + 3); np < pp;)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
*cp++ = *np++;
*cp++ = '.';
*cp = '\0';
Run msh through indent
2004-08-04 19:16:54 +00:00
for (gp = cp = space(strlen(pp) + 1); *np && *np != '/';)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
*cp++ = *np++;
*cp = '\0';
dirp = opendir(dp);
if (dirp == 0) {
DELETE(dp);
DELETE(gp);
return;
}
dname[NAME_MAX] = '\0';
Run msh through indent
2004-08-04 19:16:54 +00:00
while ((de = readdir(dirp)) != NULL) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* XXX Hmmm... What this could be? (abial) */
/*
Run msh through indent
2004-08-04 19:16:54 +00:00
if (ent[j].d_ino == 0)
continue;
*/
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
strncpy(dname, de->d_name, NAME_MAX);
Run msh through indent
2004-08-04 19:16:54 +00:00
if (dname[0] == '.')
if (*gp != '.')
continue;
for (k = 0; k < NAME_MAX; k++)
if (any(dname[k], spcl))
dname[k] |= QUOTE;
if (gmatch(dname, gp)) {
name = generate(we, pp, dname, np);
if (*np && !anys(np, spcl)) {
if (stat(name, &dbuf)) {
DELETE(name);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
continue;
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
nl = addword(name, nl);
}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
closedir(dirp);
DELETE(dp);
DELETE(gp);
}
/*
* generate a pathname as below.
* start..end1 / middle end
* the slashes come for free
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static char *generate(start1, end1, middle, end)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *start1;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *end1;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *middle, *end;
{
char *p;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *op, *xp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
p = op =
space((int) (end1 - start1) + strlen(middle) + strlen(end) + 2);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (xp = start1; xp != end1;)
*op++ = *xp++;
Run msh through indent
2004-08-04 19:16:54 +00:00
for (xp = middle; (*op++ = *xp++) != '\0';);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
op--;
Run msh through indent
2004-08-04 19:16:54 +00:00
for (xp = end; (*op++ = *xp++) != '\0';);
return (p);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int anyspcl(wb)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct wdblock *wb;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
REGISTER char **wd;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
wd = wb->w_words;
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = 0; i < wb->w_nword; i++)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (anys(spcl, *wd++))
Run msh through indent
2004-08-04 19:16:54 +00:00
return (1);
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int xstrcmp(p1, p2)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *p1, *p2;
{
Run msh through indent
2004-08-04 19:16:54 +00:00
return (strcmp(*(char **) p1, *(char **) p2));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/* -------- word.c -------- */
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct wdblock *newword(nw)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int nw;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct wdblock *wb;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
wb = (struct wdblock *) space(sizeof(*wb) + nw * sizeof(char *));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
wb->w_bsize = nw;
wb->w_nword = 0;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (wb);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct wdblock *addword(wd, wb)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *wd;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct wdblock *wb;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct wdblock *wb2;
REGISTER int nw;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (wb == NULL)
wb = newword(NSTART);
if ((nw = wb->w_nword) >= wb->w_bsize) {
wb2 = newword(nw * 2);
Run msh through indent
2004-08-04 19:16:54 +00:00
memcpy((char *) wb2->w_words, (char *) wb->w_words,
nw * sizeof(char *));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
wb2->w_nword = nw;
DELETE(wb);
wb = wb2;
}
wb->w_words[wb->w_nword++] = wd;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (wb);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
static
Run msh through indent
2004-08-04 19:16:54 +00:00
char **getwords(wb)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct wdblock *wb;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char **wd;
REGISTER int nb;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (wb == NULL)
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((char **) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (wb->w_nword == 0) {
DELETE(wb);
Run msh through indent
2004-08-04 19:16:54 +00:00
return ((char **) NULL);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
wd = (char **) space(nb = sizeof(*wd) * wb->w_nword);
Run msh through indent
2004-08-04 19:16:54 +00:00
memcpy((char *) wd, (char *) wb->w_words, nb);
DELETE(wb); /* perhaps should done by caller */
return (wd);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Reduce exported from msh applet
2005-09-22 14:38:17 +00:00
static int (*func) (char *, char *);
static int globv;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static void glob0(a0, a1, a2, a3)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *a0;
unsigned a1;
int a2;
int (*a3) (char *, char *);
{
func = a3;
globv = a2;
glob1(a0, a0 + a1 * a2);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void glob1(base, lim)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *base, *lim;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *i, *j;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int v2;
char *lptr, *hptr;
int c;
unsigned n;
v2 = globv;
Run msh through indent
2004-08-04 19:16:54 +00:00
top:
if ((n = (int) (lim - base)) <= v2)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
return;
Run msh through indent
2004-08-04 19:16:54 +00:00
n = v2 * (n / (2 * v2));
hptr = lptr = base + n;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
i = base;
Run msh through indent
2004-08-04 19:16:54 +00:00
j = lim - v2;
for (;;) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (i < lptr) {
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((c = (*func) (i, lptr)) == 0) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
glob2(i, lptr -= v2);
continue;
}
if (c < 0) {
i += v2;
continue;
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
begin:
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (j > hptr) {
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((c = (*func) (hptr, j)) == 0) {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
glob2(hptr += v2, j);
goto begin;
}
if (c > 0) {
if (i == lptr) {
glob3(i, hptr += v2, j);
i = lptr += v2;
goto begin;
}
glob2(i, j);
j -= v2;
i += v2;
continue;
}
j -= v2;
goto begin;
}
if (i == lptr) {
Run msh through indent
2004-08-04 19:16:54 +00:00
if (lptr - base >= lim - hptr) {
glob1(hptr + v2, lim);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
lim = lptr;
} else {
glob1(base, lptr);
Run msh through indent
2004-08-04 19:16:54 +00:00
base = hptr + v2;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
goto top;
}
glob3(j, lptr -= v2, i);
j = hptr -= v2;
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void glob2(i, j)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *i, *j;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *index1, *index2, c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int m;
m = globv;
index1 = i;
index2 = j;
do {
c = *index1;
*index1++ = *index2;
*index2++ = c;
Run msh through indent
2004-08-04 19:16:54 +00:00
} while (--m);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void glob3(i, j, k)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *i, *j, *k;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *index1, *index2, *index3;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int c;
int m;
m = globv;
index1 = i;
index2 = j;
index3 = k;
do {
c = *index1;
*index1++ = *index3;
*index3++ = *index2;
*index2++ = c;
Run msh through indent
2004-08-04 19:16:54 +00:00
} while (--m);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/* -------- io.c -------- */
/*
* shell IO
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int my_getc(int ec)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
if (e.linep > elinep) {
while ((c = readc()) != '\n' && c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
err("input line too long");
gflg++;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
c = readc();
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
if ((ec != '\'') && (ec != '`') && (e.iop->task != XGRAVE)) {
Run msh through indent
2004-08-04 19:16:54 +00:00
if (c == '\\') {
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
c = readc();
if (c == '\n' && ec != '\"')
Run msh through indent
2004-08-04 19:16:54 +00:00
return (my_getc(ec));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
c |= QUOTE;
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void unget(c)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int c;
{
if (e.iop >= e.iobase)
e.iop->peekc = c;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int eofc()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Run msh through indent
2004-08-04 19:16:54 +00:00
return e.iop < e.iobase || (e.iop->peekc == 0 && e.iop->prev == 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int readc()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
RCPRINTF(("READC: e.iop 0x%x, e.iobase 0x%x\n", e.iop, e.iobase));
for (; e.iop >= e.iobase; e.iop--) {
RCPRINTF(("READC: e.iop 0x%x, peekc 0x%x\n", e.iop, e.iop->peekc));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((c = e.iop->peekc) != '\0') {
e.iop->peekc = 0;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
} else {
if (e.iop->prev != 0) {
if ((c = (*e.iop->iofn) (e.iop->argp, e.iop)) != '\0') {
if (c == -1) {
e.iop++;
continue;
}
if (e.iop == iostack)
ioecho(c);
return (e.iop->prev = c);
} else if (e.iop->task == XIO && e.iop->prev != '\n') {
e.iop->prev = 0;
if (e.iop == iostack)
ioecho('\n');
return '\n';
}
}
if (e.iop->task == XIO) {
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if (multiline) {
Run msh through indent
2004-08-04 19:16:54 +00:00
return e.iop->prev = 0;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
if (interactive && e.iop == iostack + 1) {
Major rework of the directory structure and the entire build system. -Erik
2001-10-24 05:00:29 +00:00
#ifdef CONFIG_FEATURE_COMMAND_EDITING
Run msh through indent
2004-08-04 19:16:54 +00:00
current_prompt = prompt->value;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#else
Run msh through indent
2004-08-04 19:16:54 +00:00
prs(prompt->value);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#endif
Run msh through indent
2004-08-04 19:16:54 +00:00
}
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
} /* FOR */
if (e.iop >= iostack) {
RCPRINTF(("READC: return 0, e.iop 0x%x\n", e.iop));
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
}
DBGPRINTF(("READC: leave()...\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
leave();
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* NOTREACHED */
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void ioecho(c)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char c;
{
if (flag['v'])
write(2, &c, sizeof c);
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
static void pushio(struct ioarg *argp, int (*fn) (struct ioarg *))
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("PUSHIO: argp 0x%x, argp->afid 0x%x, e.iop 0x%x\n", argp,
argp->afid, e.iop));
/* Set env ptr for io source to next array spot and check for array overflow */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (++e.iop >= &iostack[NPUSH]) {
e.iop--;
err("Shell input nested too deeply");
gflg++;
return;
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
/* We did not overflow the NPUSH array spots so setup data structs */
e.iop->iofn = (int (*)(struct ioarg *, struct io *)) fn; /* Store data source func ptr */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (argp->afid != AFID_NOBUF)
Run msh through indent
2004-08-04 19:16:54 +00:00
e.iop->argp = argp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
else {
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
e.iop->argp = ioargstack + (e.iop - iostack); /* MAL - index into stack */
*e.iop->argp = *argp; /* copy data from temp area into stack spot */
/* MAL - mainbuf is for 1st data source (command line?) and all nested use a single shared buffer? */
if (e.iop == &iostack[0])
e.iop->argp->afbuf = &mainbuf;
else
e.iop->argp->afbuf = &sharedbuf;
/* MAL - if not a termimal AND (commandline OR readable file) then give it a buffer id? */
/* This line appears to be active when running scripts from command line */
if ((isatty(e.iop->argp->afile) == 0)
&& (e.iop == &iostack[0]
|| lseek(e.iop->argp->afile, 0L, 1) != -1)) {
if (++bufid == AFID_NOBUF) /* counter rollover check, AFID_NOBUF = 11111111 */
bufid = AFID_ID; /* AFID_ID = 0 */
e.iop->argp->afid = bufid; /* assign buffer id */
Run msh through indent
2004-08-04 19:16:54 +00:00
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF(("PUSHIO: iostack 0x%x, e.iop 0x%x, afbuf 0x%x\n",
iostack, e.iop, e.iop->argp->afbuf));
DBGPRINTF(("PUSHIO: mbuf 0x%x, sbuf 0x%x, bid %d, e.iop 0x%x\n",
&mainbuf, &sharedbuf, bufid, e.iop));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
e.iop->prev = ~'\n';
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
e.iop->peekc = 0;
e.iop->xchar = 0;
e.iop->nlcount = 0;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (fn == filechar || fn == linechar)
e.iop->task = XIO;
Run msh through indent
2004-08-04 19:16:54 +00:00
else if (fn == (int (*)(struct ioarg *)) gravechar
|| fn == (int (*)(struct ioarg *)) qgravechar)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
e.iop->task = XGRAVE;
else
e.iop->task = XOTHER;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
return;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static struct io *setbase(ip)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct io *ip;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct io *xp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
xp = e.iobase;
e.iobase = ip;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (xp);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Input generating functions
*/
/*
* Produce the characters of a string, then a newline, then EOF.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int nlchar(ap)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioarg *ap;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (ap->aword == NULL)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((c = *ap->aword++) == 0) {
ap->aword = NULL;
Run msh through indent
2004-08-04 19:16:54 +00:00
return ('\n');
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Given a list of words, produce the characters
* in them, with a space after each word.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int wdchar(ap)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioarg *ap;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char c;
REGISTER char **wl;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((wl = ap->awordlist) == NULL)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (*wl != NULL) {
if ((c = *(*wl)++) != 0)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c & 0177);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
ap->awordlist++;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (' ');
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
ap->awordlist = NULL;
Run msh through indent
2004-08-04 19:16:54 +00:00
return ('\n');
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Return the characters of a list of words,
* producing a space between them.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int dolchar(ap)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioarg *ap;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *wp;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((wp = *ap->awordlist++) != NULL) {
Run msh through indent
2004-08-04 19:16:54 +00:00
PUSHIO(aword, wp, *ap->awordlist == NULL ? strchar : xxchar);
return (-1);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int xxchar(ap)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioarg *ap;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (ap->aword == NULL)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((c = *ap->aword++) == '\0') {
ap->aword = NULL;
Run msh through indent
2004-08-04 19:16:54 +00:00
return (' ');
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Produce the characters from a single word (string).
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int strchar(ap)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioarg *ap;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (ap->aword == NULL || (c = *ap->aword++) == 0)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Produce quoted characters from a single word (string).
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int qstrchar(ap)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioarg *ap;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (ap->aword == NULL || (c = *ap->aword++) == 0)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (0);
return (c | QUOTE);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Return the characters from a file.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int filechar(ap)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioarg *ap;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char c;
struct iobuf *bp = ap->afbuf;
if (ap->afid != AFID_NOBUF) {
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((i = ap->afid != bp->id) || bp->bufp == bp->ebufp) {
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
if (i)
lseek(ap->afile, ap->afpos, 0);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
i = safe_read(ap->afile, bp->buf, sizeof(bp->buf));
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
if (i <= 0) {
closef(ap->afile);
return 0;
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
bp->id = ap->afid;
bp->ebufp = (bp->bufp = bp->buf) + i;
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
ap->afpos++;
return *bp->bufp++ & 0177;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Major rework of the directory structure and the entire build system. -Erik
2001-10-24 05:00:29 +00:00
#ifdef CONFIG_FEATURE_COMMAND_EDITING
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
if (interactive && isatty(ap->afile)) {
Run msh through indent
2004-08-04 19:16:54 +00:00
static char mycommand[BUFSIZ];
static int position = 0, size = 0;
while (size == 0 || position >= size) {
cmdedit_read_input(current_prompt, mycommand);
size = strlen(mycommand);
position = 0;
}
c = mycommand[position];
position++;
return (c);
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
} else
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#endif
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Patch from vodz: Changed email address cmdedit API change optimizations for traceroute and md5sum added a new shared create_icmp_socket() function
2001-07-12 20:26:32 +00:00
i = safe_read(ap->afile, &c, sizeof(c));
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
return (i == sizeof(c) ? (c & 0x7f) : (closef(ap->afile), 0));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
}
/*
* Return the characters from a here temp file.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int herechar(ap)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioarg *ap;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
char c;
if (read(ap->afile, &c, sizeof(c)) != sizeof(c)) {
close(ap->afile);
c = 0;
}
return (c);
}
/*
* Return the characters produced by a process (`...`).
* Quote them if required, and remove any trailing newline characters.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int gravechar(ap, iop)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct ioarg *ap;
struct io *iop;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
if ((c = qgravechar(ap, iop) & ~QUOTE) == '\n')
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
c = ' ';
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int qgravechar(ap, iop)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioarg *ap;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct io *iop;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
DBGPRINTF3(("QGRAVECHAR: enter, ap=0x%x, iop=0x%x\n", ap, iop));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (iop->xchar) {
if (iop->nlcount) {
iop->nlcount--;
Run msh through indent
2004-08-04 19:16:54 +00:00
return ('\n' | QUOTE);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
c = iop->xchar;
iop->xchar = 0;
} else if ((c = filechar(ap)) == '\n') {
iop->nlcount = 1;
while ((c = filechar(ap)) == '\n')
iop->nlcount++;
iop->xchar = c;
if (c == 0)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
iop->nlcount--;
c = '\n';
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c != 0 ? c | QUOTE : 0);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
/*
* Return a single command (usually the first line) from a file.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int linechar(ap)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct ioarg *ap;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((c = filechar(ap)) == '\n') {
if (!multiline) {
closef(ap->afile);
Run msh through indent
2004-08-04 19:16:54 +00:00
ap->afile = -1; /* illegal value */
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
return (c);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void prs(s)
data --> text
2005-10-17 09:48:57 +00:00
REGISTER const char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
if (*s)
write(2, s, strlen(s));
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void prn(u)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
unsigned u;
{
Patch from Jonas Holmberg, adjusted by Robert Schwebel, and later re-adjusted by me to fix msh backtick handling
2003-03-14 16:05:59 +00:00
prs(itoa(u));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void closef(i)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
if (i > 2)
close(i);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void closeall()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int u;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
for (u = NUFILE; u < NOFILE;)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
close(u++);
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/*
* remap fd into Shell's fd space
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int remap(fd)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int fd;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int map[NOFILE];
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
int newfd;
DBGPRINTF(("REMAP: fd=%d, e.iofd=%d\n", fd, e.iofd));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (fd < e.iofd) {
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = 0; i < NOFILE; i++)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
map[i] = 0;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
do {
map[fd] = 1;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
newfd = dup(fd);
fd = newfd;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
} while (fd >= 0 && fd < e.iofd);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
for (i = 0; i < NOFILE; i++)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (map[i])
close(i);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (fd < 0)
err("too many files open in shell");
}
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
Run msh through indent
2004-08-04 19:16:54 +00:00
return (fd);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static int openpipe(pv)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int *pv;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int i;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if ((i = pipe(pv)) < 0)
err("can't create pipe - try again");
Run msh through indent
2004-08-04 19:16:54 +00:00
return (i);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void closepipe(pv)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int *pv;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
if (pv != NULL) {
close(*pv++);
close(*pv);
}
}
/* -------- here.c -------- */
/*
* here documents
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static void markhere(s, iop)
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
struct ioword *iop;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct here *h, *lh;
DBGPRINTF7(("MARKHERE: enter, s=0x%x\n", s));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
h = (struct here *) space(sizeof(struct here));
if (h == 0)
return;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
h->h_tag = evalstr(s, DOSUB);
if (h->h_tag == 0)
return;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
h->h_iop = iop;
iop->io_name = 0;
h->h_next = NULL;
if (inhere == 0)
inhere = h;
else
Run msh through indent
2004-08-04 19:16:54 +00:00
for (lh = inhere; lh != NULL; lh = lh->h_next)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (lh->h_next == 0) {
lh->h_next = h;
break;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
iop->io_flag |= IOHERE | IOXHERE;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (s = h->h_tag; *s; s++)
if (*s & QUOTE) {
Run msh through indent
2004-08-04 19:16:54 +00:00
iop->io_flag &= ~IOXHERE;
*s &= ~QUOTE;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
h->h_dosub = iop->io_flag & IOXHERE;
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void gethere()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct here *h, *hp;
DBGPRINTF7(("GETHERE: enter...\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* Scan here files first leaving inhere list in place */
for (hp = h = inhere; h != NULL; hp = h, h = h->h_next)
Run msh through indent
2004-08-04 19:16:54 +00:00
readhere(&h->h_iop->io_name, h->h_tag, h->h_dosub ? 0 : '\'');
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
/* Make inhere list active - keep list intact for scraphere */
if (hp != NULL) {
Run msh through indent
2004-08-04 19:16:54 +00:00
hp->h_next = acthere;
acthere = inhere;
inhere = NULL;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void readhere(name, s, ec)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char **name;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER char *s;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int ec;
{
int tf;
char tname[30] = ".msh_XXXXXX";
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int c;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
jmp_buf ev;
Run msh through indent
2004-08-04 19:16:54 +00:00
char myline[LINELIM + 1];
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *thenext;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("READHERE: enter, name=0x%x, s=0x%x\n", name, s));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
tf = mkstemp(tname);
if (tf < 0)
return;
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
*name = strsave(tname, areanum);
if (newenv(setjmp(errpt = ev)) != 0)
unlink(tname);
else {
Run msh through indent
2004-08-04 19:16:54 +00:00
pushio(e.iop->argp, (int (*)(struct ioarg *)) e.iop->iofn);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
e.iobase = e.iop;
for (;;) {
Run msh through indent
2004-08-04 19:16:54 +00:00
if (interactive && e.iop <= iostack) {
Major rework of the directory structure and the entire build system. -Erik
2001-10-24 05:00:29 +00:00
#ifdef CONFIG_FEATURE_COMMAND_EDITING
Run msh through indent
2004-08-04 19:16:54 +00:00
current_prompt = cprompt->value;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#else
Run msh through indent
2004-08-04 19:16:54 +00:00
prs(cprompt->value);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
#endif
}
thenext = myline;
while ((c = my_getc(ec)) != '\n' && c) {
if (ec == '\'')
Run msh through indent
2004-08-04 19:16:54 +00:00
c &= ~QUOTE;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (thenext >= &myline[LINELIM]) {
c = 0;
break;
}
*thenext++ = c;
}
*thenext = 0;
if (strcmp(s, myline) == 0 || c == 0)
break;
*thenext++ = '\n';
Run msh through indent
2004-08-04 19:16:54 +00:00
write(tf, myline, (int) (thenext - myline));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
if (c == 0) {
Run msh through indent
2004-08-04 19:16:54 +00:00
prs("here document `");
prs(s);
err("' unclosed");
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
quitenv();
}
close(tf);
}
/*
* open here temp file.
* if unquoted here, expand here temp file into second temp file.
*/
Run msh through indent
2004-08-04 19:16:54 +00:00
static int herein(hname, xdoll)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
char *hname;
int xdoll;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER int hf;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int tf;
#if __GNUC__
/* Avoid longjmp clobbering */
(void) &tf;
#endif
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
if (hname == NULL)
Run msh through indent
2004-08-04 19:16:54 +00:00
return (-1);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
DBGPRINTF7(("HEREIN: hname is %s, xdoll=%d\n", hname, xdoll));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
hf = open(hname, 0);
if (hf < 0)
return (-1);
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
if (xdoll) {
char c;
char tname[30] = ".msh_XXXXXX";
jmp_buf ev;
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
tf = mkstemp(tname);
if (tf < 0)
return (-1);
if (newenv(setjmp(errpt = ev)) == 0) {
PUSHIO(afile, hf, herechar);
setbase(e.iop);
while ((c = subgetc(0, 0)) != 0) {
Run msh through indent
2004-08-04 19:16:54 +00:00
c &= ~QUOTE;
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
write(tf, &c, sizeof c);
}
quitenv();
} else
unlink(tname);
close(tf);
tf = open(tname, 0);
unlink(tname);
return (tf);
} else
return (hf);
}
Run msh through indent
2004-08-04 19:16:54 +00:00
static void scraphere()
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct here *h;
DBGPRINTF7(("SCRAPHERE: enter...\n"));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
for (h = inhere; h != NULL; h = h->h_next) {
if (h->h_iop && h->h_iop->io_name)
Run msh through indent
2004-08-04 19:16:54 +00:00
unlink(h->h_iop->io_name);
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
}
inhere = NULL;
}
/* unlink here temp files before a freearea(area) */
Run msh through indent
2004-08-04 19:16:54 +00:00
static void freehere(area)
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
int area;
{
Michael Leibow, MichaelLe at belkin.com writes: A question was posted a month ago by Mark Alamo to see if others had problems with sourcing subscripts within msh. We asked his firm to fix the msh.c bug he described because we didn't have enough time to do it ourselves. When msh.c is executing a compound statement and there is a . command to source another script file, msh.c will not execute the subscript until it's completed executing the rest of the compound statement. His example was this: Echo "Start" ; . ./subA; echo "mid" ; . ./subB ; echo "end" subA and subB execute AFTER end is printed in reverse order. The same is true if the sourced files are inside an if else fi, case esac, or any compound statement. Attached is a patch to msh.c. It fixes the problem. Cd to the root of your busybox tree and execute "patch -p1 < msh.c.patch" Unfortunately, I won't have more time to work on this so I hope that there aren't any problems! Michael Leibow Senior Software Engineer Belkin Corporation
2004-08-04 19:19:10 +00:00
REGISTER struct here *h, *hl;
DBGPRINTF6(("FREEHERE: enter, area=%d\n", area));
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
hl = NULL;
for (h = acthere; h != NULL; h = h->h_next)
if (getarea((char *) h) >= area) {
if (h->h_iop->io_name != NULL)
unlink(h->h_iop->io_name);
if (hl == NULL)
acthere = h->h_next;
else
hl->h_next = h->h_next;
} else
hl = h;
}
/*
* Copyright (c) 1987,1997, Prentice Hall
* All rights reserved.
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
*
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
* Redistribution and use of the MINIX operating system in source and
* binary forms, with or without modification, are permitted provided
* that the following conditions are met:
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
*
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
* Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
*
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
* Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials provided
* with the distribution.
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
*
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
* Neither the name of Prentice Hall nor the names of the software
* authors or contributors may be used to endorse or promote
* products derived from this software without specific prior
* written permission.
Remove trailing whitespace. Update copyright to include 2004.
2004-03-15 08:29:22 +00:00
*
This is a busybox port of the Minix shell. It is reasonably capable, handles all the basic stuff (for, case/esac, while, if/then/else), and is very small (adds just 38k on x86). It is not as rigorously correct about Bourne semantics as bash, but for most things it is quite workable. There is still some work to be done to further shrink it (it has its own globbing functions instead of using the libc ones, for example), but it is quite usable as is. -Erik
2001-06-29 04:57:14 +00:00
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS, AUTHORS, AND
* CONTRIBUTORS ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL PRENTICE HALL OR ANY AUTHORS OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
* BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
* WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
* OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
* EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
*/
Reference in New Issue Copy Permalink