storm/w3m
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix charset handling. [w3m-dev 04279]

Browse Source
This commit is contained in:
Dai Sato
2007-06-07 10:59:51 +00:00
parent 12193be8b8
commit cb4a4e90d6
3 changed files with 68 additions and 61 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
ChangeLog
View File

@@ -1,3 +1,9 @@
2007-06-07 Hironori Sakamoto <hsaka@mth.biglobe.ne.jp>
* [w3m-dev 04279] charset
* libwc/ces.h, libwc/charset.c: set charset to Shift_JIS, when locale
is japanese, jp_JP.PCK, ja_JP.IBM-932, ja_JP.IBM-943, or windows-31j.
2007-06-07 Hironori Sakamoto <hsaka@mth.biglobe.ne.jp> 2007-06-07 Hironori Sakamoto <hsaka@mth.biglobe.ne.jp>
* [w3m-dev 04278] Re: segfault on CentOS4.5/libgc6.8 * [w3m-dev 04278] Re: segfault on CentOS4.5/libgc6.8
@@ -8927,4 +8933,4 @@ a * [w3m-dev 03276] compile error on EWS4800
* release-0-2-1 * release-0-2-1
* import w3m-0.2.1 * import w3m-0.2.1
$Id: ChangeLog,v 1.1000 2007/06/07 10:54:10 inu Exp $ $Id: ChangeLog,v 1.1001 2007/06/07 10:59:51 inu Exp $

1
libwc/ces.h
View File

@@ -176,6 +176,7 @@ enum {
#define WC_CES_SHIFT_JIS (WC_CES_E_PRIV2|WC_CES_N_SHIFT_JIS) #define WC_CES_SHIFT_JIS (WC_CES_E_PRIV2|WC_CES_N_SHIFT_JIS)
#define WC_CES_CP932 WC_CES_SHIFT_JIS #define WC_CES_CP932 WC_CES_SHIFT_JIS
#define WC_CES_CP943 WC_CES_SHIFT_JIS
#define WC_CES_SHIFT_JISX0213 (WC_CES_E_PRIV2|WC_CES_N_SHIFT_JISX0213) #define WC_CES_SHIFT_JISX0213 (WC_CES_E_PRIV2|WC_CES_N_SHIFT_JISX0213)
#define WC_CES_GBK (WC_CES_E_PRIV2|WC_CES_N_GBK) #define WC_CES_GBK (WC_CES_E_PRIV2|WC_CES_N_GBK)
#define WC_CES_CP936 WC_CES_GBK #define WC_CES_CP936 WC_CES_GBK

120
libwc/charset.c
View File

@@ -32,6 +32,46 @@ static struct {
{ NULL, 0 } { NULL, 0 }
}; };
static wc_ces
wc_codepage(int n)
{
switch (n) {
case 437: return WC_CES_CP437;
case 737: return WC_CES_CP737;
case 775: return WC_CES_CP775;
case 850: return WC_CES_CP850;
case 852: return WC_CES_CP852;
case 855: return WC_CES_CP855;
case 856: return WC_CES_CP856;
case 857: return WC_CES_CP857;
case 860: return WC_CES_CP860;
case 861: return WC_CES_CP861;
case 862: return WC_CES_CP862;
case 863: return WC_CES_CP863;
case 864: return WC_CES_CP864;
case 865: return WC_CES_CP865;
case 866: return WC_CES_CP866;
case 869: return WC_CES_CP869;
case 874: return WC_CES_CP874;
case 932: return WC_CES_CP932; /* CP932 = Shift_JIS */
case 936: return WC_CES_CP936; /* CP936 = GBK > EUC_CN */
case 943: return WC_CES_CP943; /* CP943 = Shift_JIS */
case 949: return WC_CES_CP949; /* CP949 = UHC > EUC_KR */
case 950: return WC_CES_CP950; /* CP950 = Big5 */
case 1006: return WC_CES_CP1006;
case 1250: return WC_CES_CP1250;
case 1251: return WC_CES_CP1251;
case 1252: return WC_CES_CP1252;
case 1253: return WC_CES_CP1253;
case 1254: return WC_CES_CP1254;
case 1255: return WC_CES_CP1255;
case 1256: return WC_CES_CP1256;
case 1257: return WC_CES_CP1257;
case 1258: return WC_CES_CP1258;
}
return 0;
}
wc_ces wc_ces
wc_guess_charset(char *charset, wc_ces orig) wc_guess_charset(char *charset, wc_ces orig)
{ {
@@ -119,6 +159,11 @@ wc_charset_to_ces(char *charset)
if (n >= 1 && n <= 16 && n != 12) if (n >= 1 && n <= 16 && n != 12)
return (WC_CES_E_ISO_8859 | n); return (WC_CES_E_ISO_8859 | n);
return WC_CES_ISO_8859_1; return WC_CES_ISO_8859_1;
} else if (! strncmp(p, "ibm", 3)) {
p += 3;
if (*p >= '1' && *p <= '9')
return wc_codepage(atoi(p));
return wc_charset_to_ces(p);
} }
break; break;
case 'j': case 'j':
@@ -135,6 +180,10 @@ wc_charset_to_ces(char *charset)
! strncmp(p, "sjis", 4)) ! strncmp(p, "sjis", 4))
return WC_CES_SHIFT_JIS; return WC_CES_SHIFT_JIS;
break; break;
case 'p':
if (! strncmp(p, "pck", 3))
return WC_CES_SHIFT_JIS;
break;
case 'g': case 'g':
if (! strncmp(p, "gb18030", 7) || if (! strncmp(p, "gb18030", 7) ||
! strncmp(p, "gbk2k", 5)) ! strncmp(p, "gbk2k", 5))
@@ -210,58 +259,18 @@ wc_charset_to_ces(char *charset)
return WC_CES_EUC_CN; return WC_CES_EUC_CN;
if (*(p+1) != 'p') if (*(p+1) != 'p')
break; break;
n = atoi(p + 2); p += 2;
switch (n) { if (*p >= '1' && *p <= '9')
case 437: return WC_CES_CP437; return wc_codepage(atoi(p));
case 737: return WC_CES_CP737;
case 775: return WC_CES_CP775;
case 850: return WC_CES_CP850;
case 852: return WC_CES_CP852;
case 855: return WC_CES_CP855;
case 856: return WC_CES_CP856;
case 857: return WC_CES_CP857;
case 860: return WC_CES_CP860;
case 861: return WC_CES_CP861;
case 862: return WC_CES_CP862;
case 863: return WC_CES_CP863;
case 864: return WC_CES_CP864;
case 865: return WC_CES_CP865;
case 866: return WC_CES_CP866;
case 869: return WC_CES_CP869;
case 874: return WC_CES_CP874;
case 932: return WC_CES_CP932; /* CP932 = Shift_JIS */
case 936: return WC_CES_CP936; /* CP936 = GBK > EUC_CN */
case 949: return WC_CES_CP949; /* CP949 = UHC > EUC_KR */
case 950: return WC_CES_CP950; /* CP950 = Big5 */
case 1006: return WC_CES_CP1006;
case 1250: return WC_CES_CP1250;
case 1251: return WC_CES_CP1251;
case 1252: return WC_CES_CP1252;
case 1253: return WC_CES_CP1253;
case 1254: return WC_CES_CP1254;
case 1255: return WC_CES_CP1255;
case 1256: return WC_CES_CP1256;
case 1257: return WC_CES_CP1257;
case 1258: return WC_CES_CP1258;
}
break; break;
case 'w': case 'w':
if (strncmp(p, "windows", 7)) if (strncmp(p, "windows", 7))
break; break;
p += 7;
if (! strncmp(p, "31j", 3)) if (! strncmp(p, "31j", 3))
return WC_CES_CP932; return WC_CES_CP932;
n = atoi(p + 7); if (*p >= '1' && *p <= '9')
switch (n) { return wc_codepage(atoi(p));
case 1250: return WC_CES_CP1250;
case 1251: return WC_CES_CP1251;
case 1252: return WC_CES_CP1252;
case 1253: return WC_CES_CP1253;
case 1254: return WC_CES_CP1254;
case 1255: return WC_CES_CP1255;
case 1256: return WC_CES_CP1256;
case 1257: return WC_CES_CP1257;
case 1258: return WC_CES_CP1258;
}
break; break;
} }
return 0; return 0;
@@ -345,18 +354,9 @@ wc_charset_short_to_ces(char *charset)
case 'c': case 'c':
return WC_CES_ISO_2022_CN; return WC_CES_ISO_2022_CN;
case 'w': case 'w':
n = atoi(p + 1); p++;
switch (n) { if (*p >= '1' && *p <= '9')
case 1250: return WC_CES_CP1250; return wc_codepage(atoi(p));
case 1251: return WC_CES_CP1251;
case 1252: return WC_CES_CP1252;
case 1253: return WC_CES_CP1253;
case 1254: return WC_CES_CP1254;
case 1255: return WC_CES_CP1255;
case 1256: return WC_CES_CP1256;
case 1257: return WC_CES_CP1257;
case 1258: return WC_CES_CP1258;
}
break; break;
case 'r': case 'r':
return WC_CES_RAW; return WC_CES_RAW;
@@ -368,7 +368,7 @@ wc_ces
wc_locale_to_ces(char *locale) wc_locale_to_ces(char *locale)
{ {
char *p = locale; char *p = locale;
char buf[6]; char buf[8];
int n; int n;
if (*p == 'C' && *(p+1) == '\0') if (*p == 'C' && *(p+1) == '\0')
@@ -380,7 +380,7 @@ wc_locale_to_ces(char *locale)
return wc_charset_to_ces(cs); return wc_charset_to_ces(cs);
} }
#endif #endif
for (n = 0; *p && *p != '.' && n < 5; p++) { for (n = 0; *p && *p != '.' && n < 7; p++) {
if ((unsigned char)*p > 0x20) if ((unsigned char)*p > 0x20)
buf[n++] = tolower(*p); buf[n++] = tolower(*p);
} }