mirror of
https://github.com/reactos/wine.git
synced 2024-11-26 13:10:28 +00:00
a69b88b2f2
Sun Mar 15 03:46:50 1998 Dimitrie O. Paun <dimi@mail.cs.toronto.edu>
* [*/*]
Fixed some dprintf_ such that there is one and only one
new line for each dprintf and that new line occurs at the end.
Transformed some fprintfs into proper debug statements.
Removed much redundancy from most of the debug statements. The
redundancy appeared because now the component and function
name is output automatically. Most debug statements also used to
output the name of the function.
All these changes prepared the source to switch completely to
the new debugging interface.
For more info, refer to ./documentation/debug-msg
Sat Mar 14 19:45:23 1997 Andreas Mohr <100.30936@germany.net>
* [misc/shell.c] [if1632/kernel.spec]
Changed parameters of FUNC004() to fix a crash.
Not sure if this fix is correct (doc wanted).
* [windows/user.c] [if1632/user.spec] [include/user.h]
Implemented UserSeeUserDo.
* [msdos/int21.c] [include/msdos.h]
Added "GET LIST OF LISTS" (INT 21/52h).
Sat Mar 14 15:48:02 1998 Douglas Ridgway <ridgway@gmcl.com>
* [include/windows.h] [relay32/gdi32.spec] [objects/enhmetafile.c]
Beginnings of enhanced metafile support.
Fri Mar 13 20:53:09 1998 John Richardson <jrichard@zko.dec.com>
* [win32/console.c]
Restart interrupted console writes.
Fri Mar 13 18:59:24 1998 Matthew Becker <mbecker@glasscity.net>
* [*/*.c]
Updated documentation for API manpages.
* [windows/dce.c]
ReleaseDC16: Fixed cast.
* [include/windows.h] [memory/virtual.c]
VirtualQuery{Ex} should return DWORD instead of BOOL32.
Fri Mar 13 13:03:06 1998 Marcus Meissner <msmeissn@cip.informatik.uni-erlangen.de>
* [README][documentation/status/]
README updated, added lzexpand,version and multimedia
status notes to new documentation/status directory.
* [ole/*.c][if1632/typelib.spec]
Added typelib stubs, several small additions and fixes.
* [loader/pe_image.c]
Fixed a small bug (fixup_imports got passed the wrong hModule in a
remapcase).
* [loader/signal.c][if1632/signal.c][misc/winsock_dns.c]
[loader/module.c]
Fixed some recursive debugger crashes (caused by invalid FS).
* [misc/registry.c]
Two bugs fixed.
Fri Mar 13 04:55:01 1998 David Lee Lambert <lamber45@egr.msu.edu>
* [include/winnt.h] [include/winnls.h]
Moved LANG_xxx flags to winnls.h
* [include/winnls.h]
Added flags for GetDateFormat(); fixed validity of
LOCALE_SYSTEM_DEFAULT.
* [include/windows.h]
Added GetTimeFormat() prototypes.
* [ole/ole2nls.c]
Implemented ASCII date- and time-functions, using an
optimized common core; added stubs for Unicode versions;
started work on a Unicode core.
* [AUTHORS]
Added my name.
Mon Mar 9 20:10:15 1998 Eric Kohl <ekohl@abo.rhein-zeitung.de>
* [relay32/comctl32.spec] [include/imagelist.h]
[include/commctrl.h] [misc/imagelist.c] [misc/Makefile.in]
First attempt at implementing ImageLists.
Sun Mar 8 20:19:49 1998 Uwe Bonnes <bon@elektron.ikp.physik.tu-darmstadt.de>
* [files/dos_fs.c] [configure.in]
Try to get FileTimeToLocalFileTime,FileTimeToSystemTime and
SystemTimeToFileTime right.
Use timegm() where available.
* [misc/lstr.c]
Fix an off by one error in FormatMessage and handle the case
when args = NULL (used by programs to get the length of the
string).
* [win32/console.c]
Actual display a per-process Title string, better working
attempt for WriteConsole32W and ReadConsole32W.
Fri Mar 6 20:33:45 1998 Slaven Rezic <eserte@cs.tu-berlin.de>
* [include/config.h.in][configure.in][multimedia/audio.c]
[multimedia/dsound.c]
Added check for FreeBSD sound system.
Sun Mar 1 17:40:10 1998 Jason Schonberg <schon@mti.sgi.com>
* [controls/edit.c] [include/ole.h] [include/shlobj.h]
Removed final commas in enum types.
Mon Feb 23 07:52:18 1998 Luiz Otavio L. Zorzella <zorzella@nr.conexware.com>
* [multimedia/time.c]
Workaround to avoid infinite recursion inside timeGetTime.
* [multimedia/audio.c]
WODM_GETNUMDEVS and WIDM_GETNUMDEVS only return 1 now if the
SOUND_DEV can be opened, or if it's busy.
95 lines
3.0 KiB
Plaintext
95 lines
3.0 KiB
Plaintext
|
|
Console - First Pass
|
|
--------------------
|
|
|
|
Consoles are just xterms created with the -Sxxn switch.
|
|
A pty is opened and the master goes to the xterm side
|
|
and the slave is held by the wine side. The console
|
|
itself it turned into a few HANDLE32s and is set
|
|
to the STD_*_HANDLES.
|
|
|
|
It is possible to use the WriteFile and ReadFile commands
|
|
to write to a win32 console. To accomplish this, all K32OBJs
|
|
that support I/O have a read and write function pointer.
|
|
So, WriteFile calls K32OBJ_WriteFile which calls the K32OBJ's
|
|
write function pointer, which then finally calls write.
|
|
|
|
[this paragraph is now out of date]
|
|
If the command line console is to be inheirited or
|
|
a process inherits it's parents console (-- can that happen???),
|
|
the console is created at process init time via PROCESS_InheritConsole.
|
|
The 0, 1, and 2 file descriptors are duped to be the
|
|
STD_*_HANDLES in this case. Also in this case a flag is set
|
|
to indicate that the console comes from the parent process or
|
|
command line.
|
|
|
|
If a process doesn't have a console at all, it's
|
|
pdb->console is set to NULL. This helps indicate when
|
|
it is possible to create a new console (via AllocConsole).
|
|
|
|
|
|
When FreeConsole is called, all handles that the process has
|
|
open to the console are closed. Like most k32objs, if the
|
|
console's refcount reaches zero, its k32obj destroy function
|
|
is called. The destroy kills the xterm if one was open.
|
|
|
|
Also like most k32 objects, we assume that (K32OBJ) header is the
|
|
first field so the casting (from K32OBJ *to CONSOLE *)
|
|
works correctly.
|
|
|
|
FreeConsole is called on process exit (in ExitProcess) if
|
|
pdb->console is not NULL.
|
|
|
|
|
|
BUGS
|
|
----
|
|
Console processes do not inherit their parent's handles. I think
|
|
there needs to be two cases, one where they have to inherit
|
|
the stdin/stdout/stderr from unix, and one where they have to
|
|
inherit from another windows app.
|
|
|
|
SetConsoleMode -- UNIX only has ICANON and various ECHOs
|
|
to play around with for processing input. Win32 has
|
|
line-at-a-time processing, character processing, and
|
|
echo. I'm putting together an intermediate driver
|
|
that will handle this (and hopefully won't be any more
|
|
buggy then the NT4 console implementation).
|
|
|
|
|
|
================================================================
|
|
|
|
experimentation with NT4 yields that:
|
|
|
|
WriteFile
|
|
---------
|
|
o does not truncate file on 0 length write
|
|
o 0 length write or error on write changes numcharswritten to 0
|
|
o 0 length write returns TRUE
|
|
o works with console handles
|
|
|
|
_lwrite
|
|
-------
|
|
o does truncate/expand file at current position on 0 length write
|
|
o returns 0 on a zero length write
|
|
o works with console handles (typecasted)
|
|
|
|
WriteConsole
|
|
------------
|
|
o expects only console handles
|
|
|
|
|
|
SetFilePointer
|
|
--------------
|
|
o returns -1 (err 6) when used with a console handle
|
|
|
|
|
|
FreeConsole
|
|
-----------
|
|
o even when all the handles to it are freed, the win32 console
|
|
stays visible, the only way I could find to free it
|
|
was via the FreeConsole
|
|
|
|
|
|
Is it possible to interrupt win32's FileWrite? I'm not sure.
|
|
It may not be possible to interrupt any system calls.
|