add new tools

intl/chardet/tools/gen.cmd

@@ -3,3 +3,12 @@ perl geneucjp.pl     > ..\src\nsEUCJPVerifier.h
 perl geniso2022jp.pl > ..\src\nsISO2022JPVerifier.h
 perl gensjis.pl      > ..\src\nsSJISVerifier.h
 perl genutf8.pl      > ..\src\nsUTF8Verifier.h
+perl geneuckr.pl     > ..\src\nsEUCKRVerifier.h
+perl gengb2312.pl    > ..\src\nsGB2312Verifier.h
+perl genbig5.pl      > ..\src\nsBIG5Verifier.h
+perl geneuctw.pl     > ..\src\nsEUCTWVerifier.h
+perl genucs2be.pl    > ..\src\nsUCS2BEVerifier.h
+perl genucs2le.pl    > ..\src\nsUCS2LEVerifier.h
+# perl genhz.pl        > ..\src\nsHZVerifier.h
+# perl geniso2022kr.pl > ..\src\nsISO2022KRVerifier.h
+# perl geniso2022cn.pl > ..\src\nsISO2022CNVerifier.h

intl/chardet/tools/genbig5.pl

@@ -0,0 +1,37 @@
+#!/usr/local/bin/perl
+use strict;
+require "genverifier.pm";
+use genverifier;
+
+
+my(@big5_cls);
+my(@big5_st);
+my($big5_ver);
+
+
+@big5_cls = (
+ [ 0x00 , 0x00 , 0 ],
+ [ 0x1b , 0x1b , 0 ],
+ [ 0x01 , 0x3f , 1 ],
+ [ 0x40 , 0x7e , 2 ],
+ [ 0x7f , 0x7f , 1 ],
+ [ 0xff , 0xff , 0 ],
+ [ 0x80 , 0xa0 , 4 ],
+ [ 0xa1 , 0xfe , 3 ],
+);
+
+package genverifier;
+@big5_st = (
+#  0  1  2  3  4 
+   1, 0, 0, 3, 1, # state 0
+   1, 1, 1, 1, 1, # Error State - 1
+   2, 2, 2, 2, 2, # ItsMe State - 2
+   1, 1, 0, 0, 0, # state 3
+);
+
+
+$big5_ver = genverifier::GenVerifier("BIG5", "BIG5", \@big5_cls, 5,     \@big5_st);
+print $big5_ver;
+
+
+

intl/chardet/tools/geneuckr.pl

@@ -0,0 +1,37 @@
+#!/usr/local/bin/perl
+use strict;
+require "genverifier.pm";
+use genverifier;
+
+
+my(@euckr_cls);
+my(@euckr_st);
+my($euckr_ver);
+
+
+@euckr_cls = (
+ [ 0x00 , 0x00 , 0 ],
+ [ 0x1b , 0x1b , 0 ],
+ [ 0x01 , 0x7f , 1 ],
+ [ 0x80 , 0xa0 , 0 ],
+ [ 0xff , 0xff , 0 ],
+ [ 0xad , 0xaf , 3 ],
+ [ 0xc9 , 0xc9 , 3 ],
+ [ 0xa1 , 0xfe , 2 ],
+);
+
+package genverifier;
+@euckr_st = (
+#  0  1  2  3  
+   1, 0, 3, 1,  # state 0
+   1, 1, 1, 1,  # Error State - 1
+   2, 2, 2, 2,  # ItsMe State - 2
+   1, 1, 0, 0,  # state 3
+);
+
+
+$euckr_ver = genverifier::GenVerifier("EUCKR", "EUC-KR", \@euckr_cls, 4,     \@euckr_st);
+print $euckr_ver;
+
+
+

intl/chardet/tools/geneuctw.pl

@@ -0,0 +1,44 @@
+#!/usr/local/bin/perl
+use strict;
+require "genverifier.pm";
+use genverifier;
+
+
+my(@euctw_cls);
+my(@euctw_st);
+my($euctw_ver);
+
+
+@euctw_cls = (
+ [ 0x00 , 0x00 , 0 ],
+ [ 0x1b , 0x1b , 0 ],
+ [ 0x01 , 0x7f , 2 ],
+ [ 0x8e , 0x8e , 6],
+ [ 0x80 , 0xa0 , 0 ],
+ [ 0xff , 0xff , 0 ],
+ [ 0xa1 , 0xa1 , 3 ],
+ [ 0xa2 , 0xa7 , 4 ],
+ [ 0xa8 , 0xa9 , 5 ],
+ [ 0xaa , 0xc1 , 1 ],
+ [ 0xc2 , 0xc2 , 3 ],
+ [ 0xc3 , 0xc3 , 1 ],
+ [ 0xc4 , 0xfe , 3 ],
+);
+
+package genverifier;
+@euctw_st = (
+#  0  1  2  3  4  5  6
+   1, 1, 0, 3, 3, 3, 4, # state 0
+   1, 1, 1, 1, 1, 1, 1, # Error State - 1
+   2, 2, 2, 2, 2, 2, 2, # ItsMe State - 2
+   1, 0, 1, 0, 0, 0, 1, # state 3
+   1, 1, 1, 1, 5, 1, 1, # state 4
+   1, 0, 1, 0, 0, 0, 1, # state 5
+);
+
+
+$euctw_ver = genverifier::GenVerifier("EUCTW", "x-euc-tw", \@euctw_cls, 7,     \@euctw_st);
+print $euctw_ver;
+
+
+

intl/chardet/tools/gengb2312.pl

@@ -0,0 +1,36 @@
+#!/usr/local/bin/perl
+use strict;
+require "genverifier.pm";
+use genverifier;
+
+
+my(@gb2312_cls);
+my(@gb2312_st);
+my($gb2312_ver);
+
+
+@gb2312_cls = (
+ [ 0x00 , 0x00 , 0 ],
+ [ 0x1b , 0x1b , 0 ],
+ [ 0x01 , 0x7f , 1 ],
+ [ 0x80 , 0xa0 , 0 ],
+ [ 0xff , 0xff , 0 ],
+ [ 0xaa , 0xaf , 3 ],
+ [ 0xa1 , 0xfe , 2 ],
+);
+
+package genverifier;
+@gb2312_st = (
+#  0  1  2  3  
+   1, 0, 3, 1,  # state 0
+   1, 1, 1, 1,  # Error State - 1
+   2, 2, 2, 2,  # ItsMe State - 2
+   1, 1, 0, 0,  # state 3
+);
+
+
+$gb2312_ver = genverifier::GenVerifier("GB2312", "GB2312", \@gb2312_cls, 4,     \@gb2312_st);
+print $gb2312_ver;
+
+
+

intl/chardet/tools/genucs2be.pl

@@ -0,0 +1,44 @@
+#!/usr/local/bin/perl
+use strict;
+require "genverifier.pm";
+use genverifier;
+
+
+my(@ucs2be_cls);
+my(@ucs2be_st);
+my($ucs2be_ver);
+
+
+@ucs2be_cls = (
+ [ 0x00 , 0x00 , 1 ],
+ [ 0x0a , 0x0a , 4 ],
+ [ 0x0d , 0x0d , 3 ],
+ [ 0x28 , 0x2f , 0 ],
+ [ 0x34 , 0x39 , 0 ],
+ [ 0x41 , 0x4d , 0 ],
+ [ 0x80 , 0x9f , 2 ],
+ [ 0xfe , 0xfe , 5 ],
+ [ 0xff , 0xff , 6 ],
+ [ 0x00 , 0xfd , 7 ],
+);
+
+package genverifier;
+@ucs2be_st = (
+#  0  1  2  3  4  5  6  7
+   1, 5, 3, 6, 6, 7, 8, 3, # state 0
+   1, 1, 1, 1, 1, 1, 1, 1, # Error State - 1
+   2, 2, 2, 2, 2, 2, 2, 2, # ItsMe State - 2
+   4, 4, 4, 4, 4, 4, 4, 4, # state 3
+   1, 5, 3, 6, 6, 3, 3, 3, # state 4
+   4, 4, 1, 4, 4, 4, 4, 4, # state 5
+   4, 4, 4, 1, 4, 4, 4, 4, # state 6
+   4, 4, 4, 4, 4, 4, 2, 4, # state 7
+   4, 4, 4, 4, 4, 1, 4, 4, # state 8
+);
+
+
+$ucs2be_ver = genverifier::GenVerifier("UCSBE", "ISO-10646-UCS-2-BE", \@ucs2be_cls, 8,     \@ucs2be_st);
+print $ucs2be_ver;
+
+
+

intl/chardet/tools/genucs2le.pl

@@ -0,0 +1,44 @@
+#!/usr/local/bin/perl
+use strict;
+require "genverifier.pm";
+use genverifier;
+
+
+my(@ucs2le_cls);
+my(@ucs2le_st);
+my($ucs2le_ver);
+
+
+@ucs2le_cls = (
+ [ 0x00 , 0x00 , 1 ],
+ [ 0x0a , 0x0a , 4 ],
+ [ 0x0d , 0x0d , 3 ],
+ [ 0x28 , 0x2f , 0 ],
+ [ 0x34 , 0x39 , 0 ],
+ [ 0x41 , 0x4d , 0 ],
+ [ 0x80 , 0x9f , 2 ],
+ [ 0xfe , 0xfe , 5 ],
+ [ 0xff , 0xff , 6 ],
+ [ 0x00 , 0xfd , 7 ],
+);
+
+package genverifier;
+@ucs2le_st = (
+#  0  1  2  3  4  5  6  7
+   3, 3, 5, 6, 3, 8, 7, 3, # state 0
+   1, 1, 1, 1, 1, 1, 1, 1, # Error State - 1
+   2, 2, 2, 2, 2, 2, 2, 2, # ItsMe State - 2
+   1, 4, 4, 4, 4, 4, 4, 4, # state 3
+   3, 3, 5, 6, 3, 3, 3, 3, # state 4
+   4, 1, 4, 4, 4, 4, 4, 4, # state 5
+   4, 4, 4, 1, 1, 4, 4, 4, # state 6
+   1, 4, 4, 4, 4, 2, 4, 4, # state 7
+   1, 4, 4, 4, 4, 4, 1, 4, # state 8
+);
+
+
+$ucs2le_ver = genverifier::GenVerifier("UCSLE", "ISO-10646-UCS-2-LE", \@ucs2le_cls, 8,     \@ucs2le_st);
+print $ucs2le_ver;
+
+
+