Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commit5ae9d85

Browse files
committed
Add KOI8/WIN/ALT support
1 parenteb42c1c commit5ae9d85

File tree

2 files changed

+296
-37
lines changed

2 files changed

+296
-37
lines changed

‎src/backend/utils/mb/conv.c

Lines changed: 261 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -2,7 +2,7 @@
22
* conversion between client encoding and server internal encoding
33
* (currently mule internal code (mic) is used)
44
* Tatsuo Ishii
5-
* $Id: conv.c,v 1.5 1999/02/02 18:51:23 momjian Exp $
5+
* $Id: conv.c,v 1.6 1999/03/24 07:02:16 ishii Exp $
66
*/
77
#include<stdio.h>
88
#include<string.h>
@@ -588,6 +588,262 @@ mic2ascii(unsigned char *mic, unsigned char *p, int len)
588588
*p='\0';
589589
}
590590

591+
/*
592+
* Cyrillic support
593+
* currently supported Cyrillic encodings:
594+
*
595+
* KOI8-R (this is the charset for the mule internal code
596+
*for Cyrillic)
597+
* ISO-8859-5
598+
* Microsoft's CP1251(windows-1251)
599+
* Alternativny Variant (MS-DOS CP866)
600+
*/
601+
602+
/* koi2mic: KOI8-R to Mule internal code */
603+
staticvoid
604+
koi2mic(unsignedchar*l,unsignedchar*p,intlen)
605+
{
606+
latin2mic(l,p,len,LC_KOI8_R);
607+
}
608+
609+
/* mic2koi: Mule internal code to KOI8-R */
610+
staticvoid
611+
mic2koi(unsignedchar*mic,unsignedchar*p,intlen)
612+
{
613+
mic2latin(mic,p,len,LC_KOI8_R);
614+
}
615+
616+
/*
617+
* latin2mic_with_table: a generic single byte charset encoding
618+
* conversion from a local charset to the mule internal code.
619+
* with a encoding conversion table.
620+
* the table is ordered according to the local charset,
621+
* starting from 128 (0x80). each entry in the table
622+
* holds the corresponding code point for the mule internal code.
623+
*/
624+
staticvoid
625+
latin2mic_with_table(
626+
unsignedchar*l,/* local charset string (source) */
627+
unsignedchar*p,/* pointer to store mule internal code
628+
(destination) */
629+
intlen,/* length of l */
630+
intlc,/* leading character of p */
631+
unsignedchar*tab/* code conversion table */
632+
)
633+
{
634+
unsignedcharc1,c2;
635+
636+
while (len-->0&& (c1=*l++)) {
637+
if (c1<128) {
638+
*p++=c1;
639+
}else {
640+
c2=tab[c1-128];
641+
if (c2) {
642+
*p++=lc;
643+
*p++=c2;
644+
}else {
645+
*p++=' ';/* cannot convert */
646+
}
647+
}
648+
}
649+
*p='\0';
650+
}
651+
652+
/*
653+
* mic2latin_with_table: a generic single byte charset encoding
654+
* conversion from the mule internal code to a local charset
655+
* with a encoding conversion table.
656+
* the table is ordered according to the second byte of the mule
657+
* internal code starting from 128 (0x80).
658+
* each entry in the table
659+
* holds the corresponding code point for the local code.
660+
*/
661+
staticvoid
662+
mic2latin_with_table(
663+
unsignedchar*mic,/* mule internal code (source) */
664+
unsignedchar*p,/* local code (destination) */
665+
intlen,/* length of p */
666+
intlc,/* leading character */
667+
unsignedchar*tab/* code conversion table */
668+
)
669+
{
670+
671+
unsignedcharc1,c2;
672+
673+
while (len-->0&& (c1=*mic++)) {
674+
if (c1<128) {
675+
*p++=c1;
676+
}elseif (c1==lc) {
677+
c1=*mic++;
678+
len--;
679+
c2=tab[c1-128];
680+
if (c2) {
681+
*p++=c2;
682+
}else {
683+
*p++=' ';/* cannot convert */
684+
}
685+
}else {
686+
*p++=' ';/* bogus character */
687+
}
688+
}
689+
*p='\0';
690+
}
691+
692+
/* iso2mic: ISO-8859-5 to Mule internal code */
693+
staticvoid
694+
iso2mic(unsignedchar*l,unsignedchar*p,intlen)
695+
{
696+
staticchariso2koi[]= {
697+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
698+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
699+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
700+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
701+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
702+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
703+
0xe1,0xe2,0xf7,0xe7,0xe4,0xe5,0xf6,0xfa,
704+
0xe9,0xea,0xeb,0xec,0xed,0xee,0xef,0xf0,
705+
0xf2,0xf3,0xf4,0xf5,0xe6,0xe8,0xe3,0xfe,
706+
0xfb,0xfd,0xff,0xf9,0xf8,0xfc,0xe0,0xf1,
707+
0xc1,0xc2,0xd7,0xc7,0xc4,0xc5,0xd6,0xda,
708+
0xc9,0xca,0xcb,0xcc,0xcd,0xce,0xcf,0xd0,
709+
0xd2,0xd3,0xd4,0xd5,0xc6,0xc8,0xc3,0xde,
710+
0xdb,0xdd,0xdf,0xd9,0xd8,0xdc,0xc0,0xd1,
711+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
712+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00
713+
};
714+
latin2mic_with_table(l,p,len,LC_KOI8_R,iso2koi);
715+
}
716+
717+
/* mic2iso: Mule internal code to ISO8859-5 */
718+
staticvoid
719+
mic2iso(unsignedchar*mic,unsignedchar*p,intlen)
720+
{
721+
staticcharkoi2iso[]= {
722+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
723+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
724+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
725+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
726+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
727+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
728+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
729+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
730+
0xee,0xd0,0xd1,0xe6,0xd4,0xd5,0xe4,0xd3,
731+
0xe5,0xd8,0xd9,0xda,0xdb,0xdc,0xdd,0xde,
732+
0xdf,0xef,0xe0,0xe1,0xe2,0xe3,0xd6,0xd2,
733+
0xec,0xeb,0xd7,0xe8,0xed,0xe9,0xe7,0xea,
734+
0xce,0xb0,0xb1,0xc6,0xb4,0xb5,0xc4,0xb3,
735+
0xc5,0xb8,0xb9,0xba,0xbb,0xbc,0xbd,0xbe,
736+
0xbf,0xcf,0xc0,0xc1,0xc2,0xc3,0xb6,0xb2,
737+
0xcc,0xcb,0xb7,0xc8,0xcd,0xc9,0xc7,0xca
738+
};
739+
740+
mic2latin_with_table(mic,p,len,LC_KOI8_R,koi2iso);
741+
}
742+
743+
/* win2mic: CP1251 to Mule internal code */
744+
staticvoid
745+
win2mic(unsignedchar*l,unsignedchar*p,intlen)
746+
{
747+
staticcharwin2koi[]= {
748+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
749+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
750+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
751+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
752+
0x00,0x00,0x00,0x00,0x00,0xbd,0x00,0x00,
753+
0xb3,0x00,0xb4,0x00,0x00,0x00,0x00,0xb7,
754+
0x00,0x00,0xb6,0xa6,0xad,0x00,0x00,0x00,
755+
0xa3,0x00,0xa4,0x00,0x00,0x00,0x00,0xa7,
756+
0xe1,0xe2,0xf7,0xe7,0xe4,0xe5,0xf6,0xfa,
757+
0xe9,0xea,0xeb,0xec,0xed,0xee,0xef,0xf0,
758+
0xf2,0xf3,0xf4,0xf5,0xe6,0xe8,0xe3,0xfe,
759+
0xfb,0xfd,0xff,0xf9,0xf8,0xfc,0xe0,0xf1,
760+
0xc1,0xc2,0xd7,0xc7,0xc4,0xc5,0xd6,0xda,
761+
0xc9,0xca,0xcb,0xcc,0xcd,0xce,0xcf,0xd0,
762+
0xd2,0xd3,0xd4,0xd5,0xc6,0xc8,0xc3,0xde,
763+
0xdb,0xdd,0xdf,0xd9,0xd8,0xdc,0xc0,0xd1
764+
};
765+
latin2mic_with_table(l,p,len,LC_KOI8_R,win2koi);
766+
}
767+
768+
/* mic2win: Mule internal code to CP1251 */
769+
staticvoid
770+
mic2win(unsignedchar*mic,unsignedchar*p,intlen)
771+
{
772+
staticcharkoi2win[]= {
773+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
774+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
775+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
776+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
777+
0x00,0x00,0x00,0xb8,0xba,0x00,0xb3,0xbf,
778+
0x00,0x00,0x00,0x00,0x00,0xb4,0x00,0x00,
779+
0x00,0x00,0x00,0xa8,0xaa,0x00,0xb2,0xaf,
780+
0x00,0x00,0x00,0x00,0x00,0xa5,0x00,0x00,
781+
0xfe,0xe0,0xe1,0xf6,0xe4,0xe5,0xf4,0xe3,
782+
0xf5,0xe8,0xe9,0xea,0xeb,0xec,0xed,0xee,
783+
0xef,0xff,0xf0,0xf1,0xf2,0xf3,0xe6,0xe2,
784+
0xfc,0xfb,0xe7,0xf8,0xfd,0xf9,0xf7,0xfa,
785+
0xde,0xc0,0xc1,0xd6,0xc4,0xc5,0xd4,0xc3,
786+
0xd5,0xc8,0xc9,0xca,0xcb,0xcc,0xcd,0xce,
787+
0xcf,0xdf,0xd0,0xd1,0xd2,0xd3,0xc6,0xc2,
788+
0xdc,0xdb,0xc7,0xd8,0xdd,0xd9,0xd7,0xda
789+
};
790+
mic2latin_with_table(mic,p,len,LC_KOI8_R,koi2win);
791+
}
792+
793+
/* alt2mic: CP866 to Mule internal code */
794+
staticvoid
795+
alt2mic(unsignedchar*l,unsignedchar*p,intlen)
796+
{
797+
staticcharalt2koi[]= {
798+
0xe1,0xe2,0xf7,0xe7,0xe4,0xe5,0xf6,0xfa,
799+
0xe9,0xea,0xeb,0xec,0xed,0xee,0xef,0xf0,
800+
0xf2,0xf3,0xf4,0xf5,0xe6,0xe8,0xe3,0xfe,
801+
0xfb,0xfd,0xff,0xf9,0xf8,0xfc,0xe0,0xf1,
802+
0xc1,0xc2,0xd7,0xc7,0xc4,0xc5,0xd6,0xda,
803+
0xc9,0xca,0xcb,0xcc,0xcd,0xce,0xcf,0xd0,
804+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
805+
0x00,0x00,0x00,0x00,0x00,0xbd,0x00,0x00,
806+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
807+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
808+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
809+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
810+
0xd2,0xd3,0xd4,0xd5,0xc6,0xc8,0xc3,0xde,
811+
0xdb,0xdd,0xdf,0xd9,0xd8,0xdc,0xc0,0xd1,
812+
0xb3,0xa3,0xb4,0xa4,0xb7,0xa7,0x00,0x00,
813+
0xb6,0xa6,0x00,0x00,0x00,0x00,0x00,0x00
814+
};
815+
latin2mic_with_table(l,p,len,LC_KOI8_R,alt2koi);
816+
}
817+
818+
/* mic2alt: Mule internal code to CP866 */
819+
staticvoid
820+
mic2alt(unsignedchar*mic,unsignedchar*p,intlen)
821+
{
822+
staticcharkoi2alt[]= {
823+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
824+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
825+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
826+
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
827+
0x00,0x00,0x00,0xf1,0xf3,0x00,0xf9,0xf5,
828+
0x00,0x00,0x00,0x00,0x00,0xad,0x00,0x00,
829+
0x00,0x00,0x00,0xf0,0xf2,0x00,0xf8,0xf4,
830+
0x00,0x00,0x00,0x00,0x00,0xbd,0x00,0x00,
831+
0xee,0xa0,0xa1,0xe6,0xa4,0xa5,0xe4,0xa3,
832+
0xe5,0xa8,0xa9,0xaa,0xab,0xac,0xad,0xae,
833+
0xaf,0xef,0xe0,0xe1,0xe2,0xe3,0xa6,0xa2,
834+
0xec,0xeb,0xa7,0xe8,0xed,0xe9,0xe7,0xea,
835+
0x9e,0x80,0x81,0x96,0x84,0x85,0x94,0x83,
836+
0x95,0x88,0x89,0x8a,0x8b,0x8c,0x8d,0x8e,
837+
0x8f,0x9f,0x90,0x91,0x92,0x93,0x86,0x82,
838+
0x9c,0x9b,0x87,0x98,0x9d,0x99,0x97,0x9a
839+
};
840+
mic2latin_with_table(mic,p,len,LC_KOI8_R,koi2alt);
841+
}
842+
843+
/*
844+
* end of Cyrillic support
845+
*/
846+
591847
pg_encoding_conv_tblpg_conv_tbl[]= {
592848
{SQL_ASCII,"SQL_ASCII",0,ascii2mic,mic2ascii},/* SQL/ACII */
593849
{EUC_JP,"EUC_JP",0,euc_jp2mic,mic2euc_jp},/* EUC_JP */
@@ -600,7 +856,10 @@ pg_encoding_conv_tbl pg_conv_tbl[] = {
600856
{LATIN2,"LATIN2",0,latin22mic,mic2latin2},/* ISO 8859 Latin 2 */
601857
{LATIN3,"LATIN3",0,latin32mic,mic2latin3},/* ISO 8859 Latin 3 */
602858
{LATIN4,"LATIN4",0,latin42mic,mic2latin4},/* ISO 8859 Latin 4 */
603-
{LATIN5,"LATIN5",0,latin52mic,mic2latin5},/* ISO 8859 Latin 5 */
859+
{LATIN5,"LATIN5",0,iso2mic,mic2iso},/* ISO 8859 Latin 5 */
860+
{KOI8,"KOI8",0,koi2mic,mic2koi},/* KOI8-R */
861+
{WIN,"WIN",0,win2mic,mic2win},/* CP1251 */
862+
{ALT,"ALT",0,alt2mic,mic2alt},/* CP866 */
604863
{SJIS,"SJIS",1,sjis2mic,mic2sjis},/* SJIS */
605864
{BIG5,"BIG5",1,big52mic,mic2big5},/* Big5 */
606865
{-1,"",0,0,0}/* end mark */

‎src/backend/utils/mb/wchar.c

Lines changed: 35 additions & 35 deletions
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,7 @@
11
/*
22
* conversion functions between pg_wchar and multi-byte streams.
33
* Tatsuo Ishii
4-
* $Id: wchar.c,v 1.5 1999/02/02 18:51:23 momjian Exp $
4+
* $Id: wchar.c,v 1.6 1999/03/24 07:02:17 ishii Exp $
55
*/
66

77
#include"mb/pg_wchar.h"
@@ -416,40 +416,40 @@ pg_big5_mblen(const unsigned char *s)
416416
}
417417

418418
pg_wchar_tblpg_wchar_table[]= {
419-
{pg_ascii2wchar_with_len,pg_ascii_mblen},
420-
{pg_eucjp2wchar_with_len,pg_eucjp_mblen},
421-
{pg_euccn2wchar_with_len,pg_euccn_mblen},
422-
{pg_euckr2wchar_with_len,pg_euckr_mblen},
423-
{pg_euctw2wchar_with_len,pg_euctw_mblen},
424-
{pg_utf2wchar_with_len,pg_utf_mblen},
425-
{pg_mule2wchar_with_len,pg_mule_mblen},
426-
{pg_latin12wchar_with_len,pg_latin1_mblen},
427-
{pg_latin12wchar_with_len,pg_latin1_mblen},
428-
{pg_latin12wchar_with_len,pg_latin1_mblen},
429-
{pg_latin12wchar_with_len,pg_latin1_mblen},
430-
{pg_latin12wchar_with_len,pg_latin1_mblen},
431-
{0,0},
432-
{0,0},
433-
{0,0},
434-
{0,0},
435-
{0,0},
436-
{0,0},
437-
{0,0},
438-
{0,0},
439-
{0,0},
440-
{0,0},
441-
{0,0},
442-
{0,0},
443-
{0,0},
444-
{0,0},
445-
{0,0},
446-
{0,0},
447-
{0,0},
448-
{0,0},
449-
{0,0},
450-
{0,0},
451-
{0,pg_sjis_mblen},
452-
{0,pg_big5_mblen}
419+
{pg_ascii2wchar_with_len,pg_ascii_mblen},/* 0 */
420+
{pg_eucjp2wchar_with_len,pg_eucjp_mblen},/* 1 */
421+
{pg_euccn2wchar_with_len,pg_euccn_mblen},/* 2 */
422+
{pg_euckr2wchar_with_len,pg_euckr_mblen},/* 3 */
423+
{pg_euctw2wchar_with_len,pg_euctw_mblen},/* 4 */
424+
{pg_utf2wchar_with_len,pg_utf_mblen},/* 5 */
425+
{pg_mule2wchar_with_len,pg_mule_mblen},/* 6 */
426+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 7 */
427+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 8 */
428+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 9 */
429+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 10 */
430+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 11 */
431+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 12 */
432+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 13 */
433+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 14 */
434+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 15 */
435+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 16 */
436+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 17 */
437+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 18 */
438+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 19 */
439+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 20 */
440+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 21 */
441+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 22 */
442+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 23 */
443+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 24 */
444+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 25 */
445+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 26 */
446+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 27 */
447+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 28 */
448+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 29 */
449+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 30 */
450+
{pg_latin12wchar_with_len,pg_latin1_mblen},/* 31 */
451+
{0,pg_sjis_mblen},/* 32 */
452+
{0,pg_big5_mblen}/* 33 */
453453
};
454454

455455
/* returns the byte length of a word for mule internal code */

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp