Commit90a06db

committed

Fix broken GB18030 <--> UTF-8 conversion map

1 parent5eb6de5 commit90a06dbCopy full SHA for 90a06db

File tree

+126746

-126986

lines changed

+126746

-126986

lines changed

Lines changed: 24 additions & 8 deletions

Original file line number	Diff line number	Diff line change
`@@ -2,7 +2,7 @@`
`2`	`2`	`#`
`3`	`3`	`# Copyright 2002 by Bill Huang`
`4`	`4`	`#`
`5`		`-# $Id: UCS_to_GB18030.pl,v 1.1 2002/06/13 08:28:55 ishii Exp $`
	`5`	`+# $Id: UCS_to_GB18030.pl,v 1.2 2002/11/12 11:33:40 ishii Exp $`
`6`	`6`	`#`
`7`	`7`	`# Generate UTF-8 <--> GB18030 code conversion tables from`
`8`	`8`	`# map files provided by Unicode organization.`
`@@ -30,10 +30,18 @@`
`30`	`30`	`next;`
`31`	`31`	`}`
`32`	`32`	`($u,$c,$rest ) =split;`
`33`		`-$utf =hex($u);`
	`33`	`+$ucs =hex($u);`
`34`	`34`	`$code =hex($c);`
`35`		`-$count++;`
`36`		`-$array{$utf } = ($code);`
	`35`	`+if($code >= 0x80 &&$ucs >= 0x0080 ){`
	`36`	`+$utf = &ucs2utf($ucs);`
	`37`	`+if($array{$utf }ne"" ){`
	`38`	`+printfSTDERR"Warning: duplicate unicode: %04x\n",$ucs;`
	`39`	`+next;`
	`40`	`+}`
	`41`	`+$count++;`
	`42`	`+`
	`43`	`+$array{$utf } =$code;`
	`44`	`+}`
`37`	`45`	`}`
`38`	`46`	`close( FILE );`
`39`	`47`
`@@ -70,11 +78,19 @@`
`70`	`78`	`if(/^#/ ){`
`71`	`79`	`next;`
`72`	`80`	`}`
`73`		`-($u,$c,$rest ) =split;`
`74`		`-$utf =hex($u);`
	`81`	`+($c,$u,$rest ) =split;`
	`82`	`+$ucs =hex($u);`
`75`	`83`	`$code =hex($c);`
`76`		`-$count++;`
`77`		`-$array{$code } =$utf;`
	`84`	`+if($code >= 0x80 &&$ucs >= 0x0080 ){`
	`85`	`+$utf = &ucs2utf($ucs);`
	`86`	`+if($array{$code }ne"" ){`
	`87`	`+printfSTDERR"Warning: duplicate code: %04x\n",$ucs;`
	`88`	`+next;`
	`89`	`+}`
	`90`	`+$count++;`
	`91`	`+`
	`92`	`+$array{$code } =$utf;`
	`93`	`+}`
`78`	`94`	`}`
`79`	`95`	`close( FILE );`
`80`	`96`

Comments

(0)