walkero
/
fontconfig
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Guess that mac roman names with lots of high bits are actually SJIS. 

Many Japanese fonts incorrectly include names tagged as Roman encoding and
English language which are actually Japanese names in the SJIS encoding.
Guess that names with a large number of high bits set are SJIS encoded
Japanese names rather than English names.
This commit is contained in:
Keith Packard 2006-09-01 21:30:54 -07:00
parent db970d3596
commit 7295c6f5fa
1 changed files with 64 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

57
src/fcfreetype.c
View File

@ -560,6 +560,28 @@ FcFontCapabilities(FT_Face face);
#include <iconv.h>
#endif
/*
* A shift-JIS will have many high bits turned on
*/
static FcBool
FcLooksLikeSJIS (FcChar8 *string, int len)
{
int nhigh = 0, nlow = 0;
while (len-- > 0)
{
if (*string++ & 0x80) nhigh++;
else nlow++;
}
/*
* Heuristic -- if more than 1/3 of the bytes have the high-bit set,
* this is likely to be SJIS and not ROMAN
*/
if (nhigh * 2 > nlow)
return FcTrue;
return FcFalse;
}
static FcChar8 *
FcSfntNameTranscode (FT_SfntName *sname)
{
@ -579,14 +601,24 @@ FcSfntNameTranscode (FT_SfntName *sname)
return 0;
fromcode = fcFtEncoding[i].fromcode;
/*
* Many names encoded for TT_PLATFORM_MACINTOSH are broken
* in various ways. Kludge around them.
*/
if (!strcmp (fromcode, FC_ENCODING_MAC_ROMAN))
{
if (sname->language_id == TT_MAC_LANGID_ENGLISH &&
FcLooksLikeSJIS (sname->string, sname->string_len))
{
fromcode = "SJIS";
}
else if (sname->language_id >= 0x100)
{
/*
* "real" Mac language IDs are all less than 150.
* Names using one of the MS language IDs are assumed
* to use an associated encoding (Yes, this is a kludge)
*/
if (!strcmp (fromcode, FC_ENCODING_MAC_ROMAN) &&
sname->language_id >= 0x100)
{
int f;
fromcode = NULL;
@ -599,6 +631,7 @@ FcSfntNameTranscode (FT_SfntName *sname)
if (!fromcode)
return 0;
}
}
if (!strcmp (fromcode, "UCS-2BE") || !strcmp (fromcode, "UTF-16BE"))
{
FcChar8 *src = sname->string;
@ -738,10 +771,24 @@ static const FcChar8 *
FcSfntNameLanguage (FT_SfntName *sname)
{
int i;
FT_UShort platform_id = sname->platform_id;
FT_UShort language_id = sname->language_id;
/*
* Many names encoded for TT_PLATFORM_MACINTOSH are broken
* in various ways. Kludge around them.
*/
if (platform_id == TT_PLATFORM_MACINTOSH &&
sname->encoding_id == TT_MAC_ID_ROMAN &&
FcLooksLikeSJIS (sname->string, sname->string_len))
{
language_id = TT_MAC_LANGID_JAPANESE;
}
for (i = 0; i < NUM_FC_FT_LANGUAGE; i++)
if (fcFtLanguage[i].platform_id == sname->platform_id &&
if (fcFtLanguage[i].platform_id == platform_id &&
(fcFtLanguage[i].language_id == TT_LANGUAGE_DONT_CARE ||
fcFtLanguage[i].language_id == sname->language_id))
fcFtLanguage[i].language_id == language_id))
{
if (fcFtLanguage[i].lang[0] == '\0')
return NULL;