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

Commite2d088d

Browse files
committed
Allow direct conversion between EUC_JP and SJIS to improve
performance. patches submitted by Atsushi Ogawa.
1 parent1fa87fa commite2d088d

File tree

1 file changed

+201
-11
lines changed

1 file changed

+201
-11
lines changed

‎src/backend/utils/mb/conversion_procs/euc_jp_and_sjis/euc_jp_and_sjis.c

Lines changed: 201 additions & 11 deletions
Original file line numberDiff line numberDiff line change
@@ -6,7 +6,7 @@
66
* Portions Copyright (c) 1994, Regents of the University of California
77
*
88
* IDENTIFICATION
9-
* $PostgreSQL: pgsql/src/backend/utils/mb/conversion_procs/euc_jp_and_sjis/euc_jp_and_sjis.c,v 1.10 2005/06/10 16:43:56 ishii Exp $
9+
* $PostgreSQL: pgsql/src/backend/utils/mb/conversion_procs/euc_jp_and_sjis/euc_jp_and_sjis.c,v 1.11 2005/06/24 13:56:39 ishii Exp $
1010
*
1111
*-------------------------------------------------------------------------
1212
*/
@@ -58,23 +58,21 @@ static void sjis2mic(unsigned char *sjis, unsigned char *p, int len);
5858
staticvoidmic2sjis(unsignedchar*mic,unsignedchar*p,intlen);
5959
staticvoideuc_jp2mic(unsignedchar*euc,unsignedchar*p,intlen);
6060
staticvoidmic2euc_jp(unsignedchar*mic,unsignedchar*p,intlen);
61+
staticvoideuc_jp2sjis(unsignedchar*mic,unsignedchar*p,intlen);
62+
staticvoidsjis2euc_jp(unsignedchar*mic,unsignedchar*p,intlen);
6163

6264
Datum
6365
euc_jp_to_sjis(PG_FUNCTION_ARGS)
6466
{
6567
unsignedchar*src=PG_GETARG_CSTRING(2);
6668
unsignedchar*dest=PG_GETARG_CSTRING(3);
6769
intlen=PG_GETARG_INT32(4);
68-
unsignedchar*buf;
6970

7071
Assert(PG_GETARG_INT32(0)==PG_EUC_JP);
7172
Assert(PG_GETARG_INT32(1)==PG_SJIS);
7273
Assert(len >=0);
7374

74-
buf=palloc(len*ENCODING_GROWTH_RATE);
75-
euc_jp2mic(src,buf,len);
76-
mic2sjis(buf,dest,strlen(buf));
77-
pfree(buf);
75+
euc_jp2sjis(src,dest,len);
7876

7977
PG_RETURN_VOID();
8078
}
@@ -85,16 +83,12 @@ sjis_to_euc_jp(PG_FUNCTION_ARGS)
8583
unsignedchar*src=PG_GETARG_CSTRING(2);
8684
unsignedchar*dest=PG_GETARG_CSTRING(3);
8785
intlen=PG_GETARG_INT32(4);
88-
unsignedchar*buf;
8986

9087
Assert(PG_GETARG_INT32(0)==PG_SJIS);
9188
Assert(PG_GETARG_INT32(1)==PG_EUC_JP);
9289
Assert(len >=0);
9390

94-
buf=palloc(len*ENCODING_GROWTH_RATE);
95-
sjis2mic(src,buf,len);
96-
mic2euc_jp(buf,dest,strlen(buf));
97-
pfree(buf);
91+
sjis2euc_jp(src,dest,len);
9892

9993
PG_RETURN_VOID();
10094
}
@@ -454,3 +448,199 @@ mic2euc_jp(unsigned char *mic, unsigned char *p, int len)
454448
}
455449
*p='\0';
456450
}
451+
452+
/*
453+
* EUC_JP -> SJIS
454+
*/
455+
staticvoid
456+
euc_jp2sjis(unsignedchar*euc,unsignedchar*p,intlen)
457+
{
458+
intc1,
459+
c2,
460+
k;
461+
unsignedchar*euc_end=euc+len;
462+
463+
while (euc_end >=euc&& (c1=*euc++))
464+
{
465+
if(c1<0x80)
466+
{
467+
/* should be ASCII */
468+
*p++=c1;
469+
}
470+
elseif (c1==SS2)
471+
{
472+
/* hankaku kana? */
473+
*p++=*euc++;
474+
}
475+
elseif (c1==SS3)
476+
{
477+
/* JIS X0212 kanji? */
478+
c1=*euc++;
479+
c2=*euc++;
480+
k=c1 <<8 |c2;
481+
if (k >=0xf5a1)
482+
{
483+
/* UDC2 */
484+
c1-=0x54;
485+
*p++= ((c1-0xa1) >>1)+ ((c1<0xdf) ?0x81 :0xc1)+0x74;
486+
*p++=c2- ((c1&1) ? ((c2<0xe0) ?0x61 :0x60) :2);
487+
}
488+
else
489+
{
490+
inti,k2;
491+
492+
/* IBM kanji */
493+
for (i=0;;i++)
494+
{
495+
k2=ibmkanji[i].euc&0xffff;
496+
if (k2==0xffff)
497+
{
498+
*p++=PGSJISALTCODE >>8;
499+
*p++=PGSJISALTCODE&0xff;
500+
break;
501+
}
502+
if (k2==k)
503+
{
504+
k=ibmkanji[i].sjis;
505+
*p++=k >>8;
506+
*p++=k&0xff;
507+
break;
508+
}
509+
}
510+
}
511+
}
512+
else
513+
{
514+
/* JIS X0208 kanji? */
515+
c2=*euc++;
516+
k= (c1 <<8) | (c2&0xff);
517+
if (k >=0xf5a1)
518+
{
519+
/* UDC1 */
520+
c1-=0x54;
521+
*p++= ((c1-0xa1) >>1)+ ((c1<0xdf) ?0x81 :0xc1)+0x6f;
522+
}
523+
else
524+
*p++= ((c1-0xa1) >>1)+ ((c1<0xdf) ?0x81 :0xc1);
525+
*p++=c2- ((c1&1) ? ((c2<0xe0) ?0x61 :0x60) :2);
526+
}
527+
}
528+
*p='\0';
529+
}
530+
531+
/*
532+
* SJIS ---> EUC_JP
533+
*/
534+
staticvoid
535+
sjis2euc_jp(unsignedchar*sjis,unsignedchar*p,intlen)
536+
{
537+
intc1,
538+
c2,
539+
i,
540+
k,
541+
k2;
542+
unsignedchar*sjis_end=sjis+len;
543+
544+
while (sjis_end >=sjis&& (c1=*sjis++))
545+
{
546+
if(c1<0x80)
547+
{
548+
/* should be ASCII */
549+
*p++=c1;
550+
}
551+
elseif (c1 >=0xa1&&c1 <=0xdf)
552+
{
553+
/* JIS X0201 (1 byte kana) */
554+
*p++=SS2;
555+
*p++=c1;
556+
}
557+
else
558+
{
559+
/*
560+
* JIS X0208, X0212, user defined extended characters
561+
*/
562+
c2=*sjis++;
563+
k= (c1 <<8)+c2;
564+
if (k >=0xed40&&k<0xf040)
565+
{
566+
/* NEC selection IBM kanji */
567+
for (i=0;;i++)
568+
{
569+
k2=ibmkanji[i].nec;
570+
if (k2==0xffff)
571+
break;
572+
if (k2==k)
573+
{
574+
k=ibmkanji[i].sjis;
575+
c1= (k >>8)&0xff;
576+
c2=k&0xff;
577+
}
578+
}
579+
}
580+
581+
if (k<0xeb3f)
582+
{
583+
/* JIS X0208 */
584+
*p++= ((c1&0x3f) <<1)+0x9f+ (c2>0x9e);
585+
*p++=c2+ ((c2>0x9e) ?2 :0x60)+ (c2<0x80);
586+
}
587+
elseif ((k >=0xeb40&&k<0xf040)|| (k >=0xfc4c&&k <=0xfcfc))
588+
{
589+
/* NEC selection IBM kanji - Other undecided justice */
590+
*p++=PGEUCALTCODE >>8;
591+
*p++=PGEUCALTCODE&0xff;
592+
}
593+
elseif (k >=0xf040&&k<0xf540)
594+
{
595+
/*
596+
* UDC1 mapping to X0208 85 ku - 94 ku JIS code 0x7521 -
597+
* 0x7e7e EUC 0xf5a1 - 0xfefe
598+
*/
599+
c1-=0x6f;
600+
*p++= ((c1&0x3f) <<1)+0xf3+ (c2>0x9e);
601+
*p++=c2+ ((c2>0x9e) ?2 :0x60)+ (c2<0x80);
602+
}
603+
elseif (k >=0xf540&&k<0xfa40)
604+
{
605+
/*
606+
* UDC2 mapping to X0212 85 ku - 94 ku JIS code 0x7521 -
607+
* 0x7e7e EUC 0x8ff5a1 - 0x8ffefe
608+
*/
609+
*p++=SS3;
610+
c1-=0x74;
611+
*p++= ((c1&0x3f) <<1)+0xf3+ (c2>0x9e);
612+
*p++=c2+ ((c2>0x9e) ?2 :0x60)+ (c2<0x80);
613+
}
614+
elseif (k >=0xfa40)
615+
{
616+
/*
617+
* mapping IBM kanji to X0208 and X0212
618+
*
619+
*/
620+
for (i=0;;i++)
621+
{
622+
k2=ibmkanji[i].sjis;
623+
if (k2==0xffff)
624+
break;
625+
if (k2==k)
626+
{
627+
k=ibmkanji[i].euc;
628+
if (k >=0x8f0000)
629+
{
630+
*p++=SS3;
631+
*p++=0x80 | ((k&0xff00) >>8);
632+
*p++=0x80 | (k&0xff);
633+
}
634+
else
635+
{
636+
*p++=0x80 | (k >>8);
637+
*p++=0x80 | (k&0xff);
638+
}
639+
}
640+
}
641+
}
642+
}
643+
}
644+
*p='\0';
645+
}
646+

0 commit comments

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