Annotation of embedaddon/libiconv/libcharset/lib/localcharset.c, revision 1.1.1.1
1.1 misho 1: /* Determine a canonical name for the current locale's character encoding.
2:
3: Copyright (C) 2000-2006, 2008-2009 Free Software Foundation, Inc.
4:
5: This program is free software; you can redistribute it and/or modify it
6: under the terms of the GNU Library General Public License as published
7: by the Free Software Foundation; either version 2, or (at your option)
8: any later version.
9:
10: This program is distributed in the hope that it will be useful,
11: but WITHOUT ANY WARRANTY; without even the implied warranty of
12: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13: Library General Public License for more details.
14:
15: You should have received a copy of the GNU Library General Public
16: License along with this program; if not, write to the Free Software
17: Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301,
18: USA. */
19:
20: /* Written by Bruno Haible <bruno@clisp.org>. */
21:
22: #include <config.h>
23:
24: /* Specification. */
25: #include "localcharset.h"
26:
27: #include <stddef.h>
28: #include <stdio.h>
29: #include <string.h>
30: #include <stdlib.h>
31:
32: #if defined __APPLE__ && defined __MACH__ && HAVE_LANGINFO_CODESET
33: # define DARWIN7 /* Darwin 7 or newer, i.e. MacOS X 10.3 or newer */
34: #endif
35:
36: #if defined _WIN32 || defined __WIN32__
37: # define WIN32_NATIVE
38: #endif
39:
40: #if defined __EMX__
41: /* Assume EMX program runs on OS/2, even if compiled under DOS. */
42: # ifndef OS2
43: # define OS2
44: # endif
45: #endif
46:
47: #if !defined WIN32_NATIVE
48: # if HAVE_LANGINFO_CODESET
49: # include <langinfo.h>
50: # else
51: # if 0 /* see comment below */
52: # include <locale.h>
53: # endif
54: # endif
55: # ifdef __CYGWIN__
56: # define WIN32_LEAN_AND_MEAN
57: # include <windows.h>
58: # endif
59: #elif defined WIN32_NATIVE
60: # define WIN32_LEAN_AND_MEAN
61: # include <windows.h>
62: #endif
63: #if defined OS2
64: # define INCL_DOS
65: # include <os2.h>
66: #endif
67:
68: #if ENABLE_RELOCATABLE
69: # include "relocatable.h"
70: #else
71: # define relocate(pathname) (pathname)
72: #endif
73:
74: /* Get LIBDIR. */
75: #ifndef LIBDIR
76: # include "configmake.h"
77: #endif
78:
79: #if defined _WIN32 || defined __WIN32__ || defined __CYGWIN__ || defined __EMX__ || defined __DJGPP__
80: /* Win32, Cygwin, OS/2, DOS */
81: # define ISSLASH(C) ((C) == '/' || (C) == '\\')
82: #endif
83:
84: #ifndef DIRECTORY_SEPARATOR
85: # define DIRECTORY_SEPARATOR '/'
86: #endif
87:
88: #ifndef ISSLASH
89: # define ISSLASH(C) ((C) == DIRECTORY_SEPARATOR)
90: #endif
91:
92: #if HAVE_DECL_GETC_UNLOCKED
93: # undef getc
94: # define getc getc_unlocked
95: #endif
96:
97: /* The following static variable is declared 'volatile' to avoid a
98: possible multithread problem in the function get_charset_aliases. If we
99: are running in a threaded environment, and if two threads initialize
100: 'charset_aliases' simultaneously, both will produce the same value,
101: and everything will be ok if the two assignments to 'charset_aliases'
102: are atomic. But I don't know what will happen if the two assignments mix. */
103: #if __STDC__ != 1
104: # define volatile /* empty */
105: #endif
106: /* Pointer to the contents of the charset.alias file, if it has already been
107: read, else NULL. Its format is:
108: ALIAS_1 '\0' CANONICAL_1 '\0' ... ALIAS_n '\0' CANONICAL_n '\0' '\0' */
109: static const char * volatile charset_aliases;
110:
111: /* Return a pointer to the contents of the charset.alias file. */
112: static const char *
113: get_charset_aliases (void)
114: {
115: const char *cp;
116:
117: cp = charset_aliases;
118: if (cp == NULL)
119: {
120: #if !(defined DARWIN7 || defined VMS || defined WIN32_NATIVE || defined __CYGWIN__)
121: FILE *fp;
122: const char *dir;
123: const char *base = "charset.alias";
124: char *file_name;
125:
126: /* Make it possible to override the charset.alias location. This is
127: necessary for running the testsuite before "make install". */
128: dir = getenv ("CHARSETALIASDIR");
129: if (dir == NULL || dir[0] == '\0')
130: dir = relocate (LIBDIR);
131:
132: /* Concatenate dir and base into freshly allocated file_name. */
133: {
134: size_t dir_len = strlen (dir);
135: size_t base_len = strlen (base);
136: int add_slash = (dir_len > 0 && !ISSLASH (dir[dir_len - 1]));
137: file_name = (char *) malloc (dir_len + add_slash + base_len + 1);
138: if (file_name != NULL)
139: {
140: memcpy (file_name, dir, dir_len);
141: if (add_slash)
142: file_name[dir_len] = DIRECTORY_SEPARATOR;
143: memcpy (file_name + dir_len + add_slash, base, base_len + 1);
144: }
145: }
146:
147: if (file_name == NULL || (fp = fopen (file_name, "r")) == NULL)
148: /* Out of memory or file not found, treat it as empty. */
149: cp = "";
150: else
151: {
152: /* Parse the file's contents. */
153: char *res_ptr = NULL;
154: size_t res_size = 0;
155:
156: for (;;)
157: {
158: int c;
159: char buf1[50+1];
160: char buf2[50+1];
161: size_t l1, l2;
162: char *old_res_ptr;
163:
164: c = getc (fp);
165: if (c == EOF)
166: break;
167: if (c == '\n' || c == ' ' || c == '\t')
168: continue;
169: if (c == '#')
170: {
171: /* Skip comment, to end of line. */
172: do
173: c = getc (fp);
174: while (!(c == EOF || c == '\n'));
175: if (c == EOF)
176: break;
177: continue;
178: }
179: ungetc (c, fp);
180: if (fscanf (fp, "%50s %50s", buf1, buf2) < 2)
181: break;
182: l1 = strlen (buf1);
183: l2 = strlen (buf2);
184: old_res_ptr = res_ptr;
185: if (res_size == 0)
186: {
187: res_size = l1 + 1 + l2 + 1;
188: res_ptr = (char *) malloc (res_size + 1);
189: }
190: else
191: {
192: res_size += l1 + 1 + l2 + 1;
193: res_ptr = (char *) realloc (res_ptr, res_size + 1);
194: }
195: if (res_ptr == NULL)
196: {
197: /* Out of memory. */
198: res_size = 0;
199: if (old_res_ptr != NULL)
200: free (old_res_ptr);
201: break;
202: }
203: strcpy (res_ptr + res_size - (l2 + 1) - (l1 + 1), buf1);
204: strcpy (res_ptr + res_size - (l2 + 1), buf2);
205: }
206: fclose (fp);
207: if (res_size == 0)
208: cp = "";
209: else
210: {
211: *(res_ptr + res_size) = '\0';
212: cp = res_ptr;
213: }
214: }
215:
216: if (file_name != NULL)
217: free (file_name);
218:
219: #else
220:
221: # if defined DARWIN7
222: /* To avoid the trouble of installing a file that is shared by many
223: GNU packages -- many packaging systems have problems with this --,
224: simply inline the aliases here. */
225: cp = "ISO8859-1" "\0" "ISO-8859-1" "\0"
226: "ISO8859-2" "\0" "ISO-8859-2" "\0"
227: "ISO8859-4" "\0" "ISO-8859-4" "\0"
228: "ISO8859-5" "\0" "ISO-8859-5" "\0"
229: "ISO8859-7" "\0" "ISO-8859-7" "\0"
230: "ISO8859-9" "\0" "ISO-8859-9" "\0"
231: "ISO8859-13" "\0" "ISO-8859-13" "\0"
232: "ISO8859-15" "\0" "ISO-8859-15" "\0"
233: "KOI8-R" "\0" "KOI8-R" "\0"
234: "KOI8-U" "\0" "KOI8-U" "\0"
235: "CP866" "\0" "CP866" "\0"
236: "CP949" "\0" "CP949" "\0"
237: "CP1131" "\0" "CP1131" "\0"
238: "CP1251" "\0" "CP1251" "\0"
239: "eucCN" "\0" "GB2312" "\0"
240: "GB2312" "\0" "GB2312" "\0"
241: "eucJP" "\0" "EUC-JP" "\0"
242: "eucKR" "\0" "EUC-KR" "\0"
243: "Big5" "\0" "BIG5" "\0"
244: "Big5HKSCS" "\0" "BIG5-HKSCS" "\0"
245: "GBK" "\0" "GBK" "\0"
246: "GB18030" "\0" "GB18030" "\0"
247: "SJIS" "\0" "SHIFT_JIS" "\0"
248: "ARMSCII-8" "\0" "ARMSCII-8" "\0"
249: "PT154" "\0" "PT154" "\0"
250: /*"ISCII-DEV" "\0" "?" "\0"*/
251: "*" "\0" "UTF-8" "\0";
252: # endif
253:
254: # if defined VMS
255: /* To avoid the troubles of an extra file charset.alias_vms in the
256: sources of many GNU packages, simply inline the aliases here. */
257: /* The list of encodings is taken from the OpenVMS 7.3-1 documentation
258: "Compaq C Run-Time Library Reference Manual for OpenVMS systems"
259: section 10.7 "Handling Different Character Sets". */
260: cp = "ISO8859-1" "\0" "ISO-8859-1" "\0"
261: "ISO8859-2" "\0" "ISO-8859-2" "\0"
262: "ISO8859-5" "\0" "ISO-8859-5" "\0"
263: "ISO8859-7" "\0" "ISO-8859-7" "\0"
264: "ISO8859-8" "\0" "ISO-8859-8" "\0"
265: "ISO8859-9" "\0" "ISO-8859-9" "\0"
266: /* Japanese */
267: "eucJP" "\0" "EUC-JP" "\0"
268: "SJIS" "\0" "SHIFT_JIS" "\0"
269: "DECKANJI" "\0" "DEC-KANJI" "\0"
270: "SDECKANJI" "\0" "EUC-JP" "\0"
271: /* Chinese */
272: "eucTW" "\0" "EUC-TW" "\0"
273: "DECHANYU" "\0" "DEC-HANYU" "\0"
274: "DECHANZI" "\0" "GB2312" "\0"
275: /* Korean */
276: "DECKOREAN" "\0" "EUC-KR" "\0";
277: # endif
278:
279: # if defined WIN32_NATIVE || defined __CYGWIN__
280: /* To avoid the troubles of installing a separate file in the same
281: directory as the DLL and of retrieving the DLL's directory at
282: runtime, simply inline the aliases here. */
283:
284: cp = "CP936" "\0" "GBK" "\0"
285: "CP1361" "\0" "JOHAB" "\0"
286: "CP20127" "\0" "ASCII" "\0"
287: "CP20866" "\0" "KOI8-R" "\0"
288: "CP20936" "\0" "GB2312" "\0"
289: "CP21866" "\0" "KOI8-RU" "\0"
290: "CP28591" "\0" "ISO-8859-1" "\0"
291: "CP28592" "\0" "ISO-8859-2" "\0"
292: "CP28593" "\0" "ISO-8859-3" "\0"
293: "CP28594" "\0" "ISO-8859-4" "\0"
294: "CP28595" "\0" "ISO-8859-5" "\0"
295: "CP28596" "\0" "ISO-8859-6" "\0"
296: "CP28597" "\0" "ISO-8859-7" "\0"
297: "CP28598" "\0" "ISO-8859-8" "\0"
298: "CP28599" "\0" "ISO-8859-9" "\0"
299: "CP28605" "\0" "ISO-8859-15" "\0"
300: "CP38598" "\0" "ISO-8859-8" "\0"
301: "CP51932" "\0" "EUC-JP" "\0"
302: "CP51936" "\0" "GB2312" "\0"
303: "CP51949" "\0" "EUC-KR" "\0"
304: "CP51950" "\0" "EUC-TW" "\0"
305: "CP54936" "\0" "GB18030" "\0"
306: "CP65001" "\0" "UTF-8" "\0";
307: # endif
308: #endif
309:
310: charset_aliases = cp;
311: }
312:
313: return cp;
314: }
315:
316: /* Determine the current locale's character encoding, and canonicalize it
317: into one of the canonical names listed in config.charset.
318: The result must not be freed; it is statically allocated.
319: If the canonical name cannot be determined, the result is a non-canonical
320: name. */
321:
322: #ifdef STATIC
323: STATIC
324: #endif
325: const char *
326: locale_charset (void)
327: {
328: const char *codeset;
329: const char *aliases;
330:
331: #if !(defined WIN32_NATIVE || defined OS2)
332:
333: # if HAVE_LANGINFO_CODESET
334:
335: /* Most systems support nl_langinfo (CODESET) nowadays. */
336: codeset = nl_langinfo (CODESET);
337:
338: # ifdef __CYGWIN__
339: /* Cygwin 2006 does not have locales. nl_langinfo (CODESET) always
340: returns "US-ASCII". As long as this is not fixed, return the suffix
341: of the locale name from the environment variables (if present) or
342: the codepage as a number. */
343: if (codeset != NULL && strcmp (codeset, "US-ASCII") == 0)
344: {
345: const char *locale;
346: static char buf[2 + 10 + 1];
347:
348: locale = getenv ("LC_ALL");
349: if (locale == NULL || locale[0] == '\0')
350: {
351: locale = getenv ("LC_CTYPE");
352: if (locale == NULL || locale[0] == '\0')
353: locale = getenv ("LANG");
354: }
355: if (locale != NULL && locale[0] != '\0')
356: {
357: /* If the locale name contains an encoding after the dot, return
358: it. */
359: const char *dot = strchr (locale, '.');
360:
361: if (dot != NULL)
362: {
363: const char *modifier;
364:
365: dot++;
366: /* Look for the possible @... trailer and remove it, if any. */
367: modifier = strchr (dot, '@');
368: if (modifier == NULL)
369: return dot;
370: if (modifier - dot < sizeof (buf))
371: {
372: memcpy (buf, dot, modifier - dot);
373: buf [modifier - dot] = '\0';
374: return buf;
375: }
376: }
377: }
378:
379: /* Woe32 has a function returning the locale's codepage as a number. */
380: sprintf (buf, "CP%u", GetACP ());
381: codeset = buf;
382: }
383: # endif
384:
385: # else
386:
387: /* On old systems which lack it, use setlocale or getenv. */
388: const char *locale = NULL;
389:
390: /* But most old systems don't have a complete set of locales. Some
391: (like SunOS 4 or DJGPP) have only the C locale. Therefore we don't
392: use setlocale here; it would return "C" when it doesn't support the
393: locale name the user has set. */
394: # if 0
395: locale = setlocale (LC_CTYPE, NULL);
396: # endif
397: if (locale == NULL || locale[0] == '\0')
398: {
399: locale = getenv ("LC_ALL");
400: if (locale == NULL || locale[0] == '\0')
401: {
402: locale = getenv ("LC_CTYPE");
403: if (locale == NULL || locale[0] == '\0')
404: locale = getenv ("LANG");
405: }
406: }
407:
408: /* On some old systems, one used to set locale = "iso8859_1". On others,
409: you set it to "language_COUNTRY.charset". In any case, we resolve it
410: through the charset.alias file. */
411: codeset = locale;
412:
413: # endif
414:
415: #elif defined WIN32_NATIVE
416:
417: static char buf[2 + 10 + 1];
418:
419: /* Woe32 has a function returning the locale's codepage as a number. */
420: sprintf (buf, "CP%u", GetACP ());
421: codeset = buf;
422:
423: #elif defined OS2
424:
425: const char *locale;
426: static char buf[2 + 10 + 1];
427: ULONG cp[3];
428: ULONG cplen;
429:
430: /* Allow user to override the codeset, as set in the operating system,
431: with standard language environment variables. */
432: locale = getenv ("LC_ALL");
433: if (locale == NULL || locale[0] == '\0')
434: {
435: locale = getenv ("LC_CTYPE");
436: if (locale == NULL || locale[0] == '\0')
437: locale = getenv ("LANG");
438: }
439: if (locale != NULL && locale[0] != '\0')
440: {
441: /* If the locale name contains an encoding after the dot, return it. */
442: const char *dot = strchr (locale, '.');
443:
444: if (dot != NULL)
445: {
446: const char *modifier;
447:
448: dot++;
449: /* Look for the possible @... trailer and remove it, if any. */
450: modifier = strchr (dot, '@');
451: if (modifier == NULL)
452: return dot;
453: if (modifier - dot < sizeof (buf))
454: {
455: memcpy (buf, dot, modifier - dot);
456: buf [modifier - dot] = '\0';
457: return buf;
458: }
459: }
460:
461: /* Resolve through the charset.alias file. */
462: codeset = locale;
463: }
464: else
465: {
466: /* OS/2 has a function returning the locale's codepage as a number. */
467: if (DosQueryCp (sizeof (cp), cp, &cplen))
468: codeset = "";
469: else
470: {
471: sprintf (buf, "CP%u", cp[0]);
472: codeset = buf;
473: }
474: }
475:
476: #endif
477:
478: if (codeset == NULL)
479: /* The canonical name cannot be determined. */
480: codeset = "";
481:
482: /* Resolve alias. */
483: for (aliases = get_charset_aliases ();
484: *aliases != '\0';
485: aliases += strlen (aliases) + 1, aliases += strlen (aliases) + 1)
486: if (strcmp (codeset, aliases) == 0
487: || (aliases[0] == '*' && aliases[1] == '\0'))
488: {
489: codeset = aliases + strlen (aliases) + 1;
490: break;
491: }
492:
493: /* Don't return an empty string. GNU libc and GNU libiconv interpret
494: the empty string as denoting "the locale's character encoding",
495: thus GNU libiconv would call this function a second time. */
496: if (codeset[0] == '\0')
497: codeset = "ASCII";
498:
499: return codeset;
500: }
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>
![[BACK]](/icons/cvsweb/back.gif){kind=icon}
Return to localcharset.c CVS log ![[TXT]](/icons/cvsweb/text.gif){kind=icon}
![[DIR]](/icons/cvsweb/dir.gif){kind=icon}
Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / libiconv / libcharset / lib