1 /* Determine a canonical name for the current locale's character encoding.
2
3 Copyright (C) 2000-2006, 2008-2018 Free Software Foundation, Inc.
4
5 This program is free software; you can redistribute it and/or modify
6 it under the terms of the GNU General Public License as published by
7 the Free Software Foundation; either version 3, 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
13 GNU General Public License for more details.
14
15 You should have received a copy of the GNU General Public License along
16 with this program; if not, see <https://www.gnu.org/licenses/>. */
17
18 /* Written by Bruno Haible <bruno@clisp.org>. */
19
20 #include <config.h>
21
22 /* Specification. */
23 #include "localcharset.h"
24
25 #include <stddef.h>
26 #include <stdio.h>
27 #include <string.h>
28 #include <stdlib.h>
29
30 #if defined __APPLE__ && defined __MACH__ && HAVE_LANGINFO_CODESET
31 # define DARWIN7 /* Darwin 7 or newer, i.e. Mac OS X 10.3 or newer */
32 #endif
33
34 #if defined _WIN32 && !defined __CYGWIN__
35 # define WINDOWS_NATIVE
36 # include <locale.h>
37 #endif
38
39 #if defined __EMX__
40 /* Assume EMX program runs on OS/2, even if compiled under DOS. */
41 # ifndef OS2
42 # define OS2
43 # endif
44 #endif
45
46 #if !defined WINDOWS_NATIVE
47 # if HAVE_LANGINFO_CODESET
48 # include <langinfo.h>
49 # else
50 # if 0 /* see comment regarding use of setlocale(), below */
51 # include <locale.h>
52 # endif
53 # endif
54 # ifdef __CYGWIN__
55 # define WIN32_LEAN_AND_MEAN
56 # include <windows.h>
57 # endif
58 #elif defined WINDOWS_NATIVE
59 # define WIN32_LEAN_AND_MEAN
60 # include <windows.h>
61 #endif
62 #if defined OS2
63 # define INCL_DOS
64 # include <os2.h>
65 #endif
66
67 /* For MB_CUR_MAX_L */
68 #if defined DARWIN7
69 # include <xlocale.h>
70 #endif
71
72
73 #if HAVE_LANGINFO_CODESET || defined WINDOWS_NATIVE || defined OS2
74
75 /* On these platforms, we use a mapping from non-canonical encoding name
76 to GNU canonical encoding name. */
77
78 /* With glibc-2.1 or newer, we don't need any canonicalization,
79 because glibc has iconv and both glibc and libiconv support all
80 GNU canonical names directly. */
81 # if !((defined __GNU_LIBRARY__ && __GLIBC__ >= 2) || defined __UCLIBC__)
82
83 struct table_entry
84 {
85 const char alias[11+1];
86 const char canonical[11+1];
87 };
88
89 /* Table of platform-dependent mappings, sorted in ascending order. */
90 static const struct table_entry alias_table[] =
91 {
92 # if defined __FreeBSD__ /* FreeBSD */
93 /*{ "ARMSCII-8", "ARMSCII-8" },*/
94 { "Big5", "BIG5" },
95 { "C", "ASCII" },
96 /*{ "CP1131", "CP1131" },*/
97 /*{ "CP1251", "CP1251" },*/
98 /*{ "CP866", "CP866" },*/
99 /*{ "GB18030", "GB18030" },*/
100 /*{ "GB2312", "GB2312" },*/
101 /*{ "GBK", "GBK" },*/
102 /*{ "ISCII-DEV", "?" },*/
103 { "ISO8859-1", "ISO-8859-1" },
104 { "ISO8859-13", "ISO-8859-13" },
105 { "ISO8859-15", "ISO-8859-15" },
106 { "ISO8859-2", "ISO-8859-2" },
107 { "ISO8859-5", "ISO-8859-5" },
108 { "ISO8859-7", "ISO-8859-7" },
109 { "ISO8859-9", "ISO-8859-9" },
110 /*{ "KOI8-R", "KOI8-R" },*/
111 /*{ "KOI8-U", "KOI8-U" },*/
112 { "SJIS", "SHIFT_JIS" },
113 { "US-ASCII", "ASCII" },
114 { "eucCN", "GB2312" },
115 { "eucJP", "EUC-JP" },
116 { "eucKR", "EUC-KR" }
117 # define alias_table_defined
118 # endif
119 # if defined __NetBSD__ /* NetBSD */
120 { "646", "ASCII" },
121 /*{ "ARMSCII-8", "ARMSCII-8" },*/
122 /*{ "BIG5", "BIG5" },*/
123 { "Big5-HKSCS", "BIG5-HKSCS" },
124 /*{ "CP1251", "CP1251" },*/
125 /*{ "CP866", "CP866" },*/
126 /*{ "GB18030", "GB18030" },*/
127 /*{ "GB2312", "GB2312" },*/
128 { "ISO8859-1", "ISO-8859-1" },
129 { "ISO8859-13", "ISO-8859-13" },
130 { "ISO8859-15", "ISO-8859-15" },
131 { "ISO8859-2", "ISO-8859-2" },
132 { "ISO8859-4", "ISO-8859-4" },
133 { "ISO8859-5", "ISO-8859-5" },
134 { "ISO8859-7", "ISO-8859-7" },
135 /*{ "KOI8-R", "KOI8-R" },*/
136 /*{ "KOI8-U", "KOI8-U" },*/
137 /*{ "PT154", "PT154" },*/
138 { "SJIS", "SHIFT_JIS" },
139 { "eucCN", "GB2312" },
140 { "eucJP", "EUC-JP" },
141 { "eucKR", "EUC-KR" },
142 { "eucTW", "EUC-TW" }
143 # define alias_table_defined
144 # endif
145 # if defined __OpenBSD__ /* OpenBSD */
146 { "646", "ASCII" },
147 { "ISO8859-1", "ISO-8859-1" },
148 { "ISO8859-13", "ISO-8859-13" },
149 { "ISO8859-15", "ISO-8859-15" },
150 { "ISO8859-2", "ISO-8859-2" },
151 { "ISO8859-4", "ISO-8859-4" },
152 { "ISO8859-5", "ISO-8859-5" },
153 { "ISO8859-7", "ISO-8859-7" }
154 # define alias_table_defined
155 # endif
156 # if defined __APPLE__ && defined __MACH__ /* Mac OS X */
157 /* Darwin 7.5 has nl_langinfo(CODESET), but sometimes its value is
158 useless:
159 - It returns the empty string when LANG is set to a locale of the
160 form ll_CC, although ll_CC/LC_CTYPE is a symlink to an UTF-8
161 LC_CTYPE file.
162 - The environment variables LANG, LC_CTYPE, LC_ALL are not set by
163 the system; nl_langinfo(CODESET) returns "US-ASCII" in this case.
164 - The documentation says:
165 "... all code that calls BSD system routines should ensure
166 that the const *char parameters of these routines are in UTF-8
167 encoding. All BSD system functions expect their string
168 parameters to be in UTF-8 encoding and nothing else."
169 It also says
170 "An additional caveat is that string parameters for files,
171 paths, and other file-system entities must be in canonical
172 UTF-8. In a canonical UTF-8 Unicode string, all decomposable
173 characters are decomposed ..."
174 but this is not true: You can pass non-decomposed UTF-8 strings
175 to file system functions, and it is the OS which will convert
176 them to decomposed UTF-8 before accessing the file system.
177 - The Apple Terminal application displays UTF-8 by default.
178 - However, other applications are free to use different encodings:
179 - xterm uses ISO-8859-1 by default.
180 - TextEdit uses MacRoman by default.
181 We prefer UTF-8 over decomposed UTF-8-MAC because one should
182 minimize the use of decomposed Unicode. Unfortunately, through the
183 Darwin file system, decomposed UTF-8 strings are leaked into user
184 space nevertheless.
185 Then there are also the locales with encodings other than US-ASCII
186 and UTF-8. These locales can be occasionally useful to users (e.g.
187 when grepping through ISO-8859-1 encoded text files), when all their
188 file names are in US-ASCII.
189 */
190 { "ARMSCII-8", "ARMSCII-8" },
191 { "Big5", "BIG5" },
192 { "Big5HKSCS", "BIG5-HKSCS" },
193 { "CP1131", "CP1131" },
194 { "CP1251", "CP1251" },
195 { "CP866", "CP866" },
196 { "CP949", "CP949" },
197 { "GB18030", "GB18030" },
198 { "GB2312", "GB2312" },
199 { "GBK", "GBK" },
200 /*{ "ISCII-DEV", "?" },*/
201 { "ISO8859-1", "ISO-8859-1" },
202 { "ISO8859-13", "ISO-8859-13" },
203 { "ISO8859-15", "ISO-8859-15" },
204 { "ISO8859-2", "ISO-8859-2" },
205 { "ISO8859-4", "ISO-8859-4" },
206 { "ISO8859-5", "ISO-8859-5" },
207 { "ISO8859-7", "ISO-8859-7" },
208 { "ISO8859-9", "ISO-8859-9" },
209 { "KOI8-R", "KOI8-R" },
210 { "KOI8-U", "KOI8-U" },
211 { "PT154", "PT154" },
212 { "SJIS", "SHIFT_JIS" },
213 { "eucCN", "GB2312" },
214 { "eucJP", "EUC-JP" },
215 { "eucKR", "EUC-KR" }
216 # define alias_table_defined
217 # endif
218 # if defined _AIX /* AIX */
219 /*{ "GBK", "GBK" },*/
220 { "IBM-1046", "CP1046" },
221 { "IBM-1124", "CP1124" },
222 { "IBM-1129", "CP1129" },
223 { "IBM-1252", "CP1252" },
224 { "IBM-850", "CP850" },
225 { "IBM-856", "CP856" },
226 { "IBM-921", "ISO-8859-13" },
227 { "IBM-922", "CP922" },
228 { "IBM-932", "CP932" },
229 { "IBM-943", "CP943" },
230 { "IBM-eucCN", "GB2312" },
231 { "IBM-eucJP", "EUC-JP" },
232 { "IBM-eucKR", "EUC-KR" },
233 { "IBM-eucTW", "EUC-TW" },
234 { "ISO8859-1", "ISO-8859-1" },
235 { "ISO8859-15", "ISO-8859-15" },
236 { "ISO8859-2", "ISO-8859-2" },
237 { "ISO8859-5", "ISO-8859-5" },
238 { "ISO8859-6", "ISO-8859-6" },
239 { "ISO8859-7", "ISO-8859-7" },
240 { "ISO8859-8", "ISO-8859-8" },
241 { "ISO8859-9", "ISO-8859-9" },
242 { "TIS-620", "TIS-620" },
243 /*{ "UTF-8", "UTF-8" },*/
244 { "big5", "BIG5" }
245 # define alias_table_defined
246 # endif
247 # if defined __hpux /* HP-UX */
248 { "SJIS", "SHIFT_JIS" },
249 { "arabic8", "HP-ARABIC8" },
250 { "big5", "BIG5" },
251 { "cp1251", "CP1251" },
252 { "eucJP", "EUC-JP" },
253 { "eucKR", "EUC-KR" },
254 { "eucTW", "EUC-TW" },
255 { "gb18030", "GB18030" },
256 { "greek8", "HP-GREEK8" },
257 { "hebrew8", "HP-HEBREW8" },
258 { "hkbig5", "BIG5-HKSCS" },
259 { "hp15CN", "GB2312" },
260 { "iso88591", "ISO-8859-1" },
261 { "iso885913", "ISO-8859-13" },
262 { "iso885915", "ISO-8859-15" },
263 { "iso88592", "ISO-8859-2" },
264 { "iso88594", "ISO-8859-4" },
265 { "iso88595", "ISO-8859-5" },
266 { "iso88596", "ISO-8859-6" },
267 { "iso88597", "ISO-8859-7" },
268 { "iso88598", "ISO-8859-8" },
269 { "iso88599", "ISO-8859-9" },
270 { "kana8", "HP-KANA8" },
271 { "koi8r", "KOI8-R" },
272 { "roman8", "HP-ROMAN8" },
273 { "tis620", "TIS-620" },
274 { "turkish8", "HP-TURKISH8" },
275 { "utf8", "UTF-8" }
276 # define alias_table_defined
277 # endif
278 # if defined __sgi /* IRIX */
279 { "ISO8859-1", "ISO-8859-1" },
280 { "ISO8859-15", "ISO-8859-15" },
281 { "ISO8859-2", "ISO-8859-2" },
282 { "ISO8859-5", "ISO-8859-5" },
283 { "ISO8859-7", "ISO-8859-7" },
284 { "ISO8859-9", "ISO-8859-9" },
285 { "eucCN", "GB2312" },
286 { "eucJP", "EUC-JP" },
287 { "eucKR", "EUC-KR" },
288 { "eucTW", "EUC-TW" }
289 # define alias_table_defined
290 # endif
291 # if defined __osf__ /* OSF/1 */
292 /*{ "GBK", "GBK" },*/
293 { "ISO8859-1", "ISO-8859-1" },
294 { "ISO8859-15", "ISO-8859-15" },
295 { "ISO8859-2", "ISO-8859-2" },
296 { "ISO8859-4", "ISO-8859-4" },
297 { "ISO8859-5", "ISO-8859-5" },
298 { "ISO8859-7", "ISO-8859-7" },
299 { "ISO8859-8", "ISO-8859-8" },
300 { "ISO8859-9", "ISO-8859-9" },
301 { "KSC5601", "CP949" },
302 { "SJIS", "SHIFT_JIS" },
303 { "TACTIS", "TIS-620" },
304 /*{ "UTF-8", "UTF-8" },*/
305 { "big5", "BIG5" },
306 { "cp850", "CP850" },
307 { "dechanyu", "DEC-HANYU" },
308 { "dechanzi", "GB2312" },
309 { "deckanji", "DEC-KANJI" },
310 { "deckorean", "EUC-KR" },
311 { "eucJP", "EUC-JP" },
312 { "eucKR", "EUC-KR" },
313 { "eucTW", "EUC-TW" },
314 { "sdeckanji", "EUC-JP" }
315 # define alias_table_defined
316 # endif
317 # if defined __sun /* Solaris */
318 { "5601", "EUC-KR" },
319 { "646", "ASCII" },
320 /*{ "BIG5", "BIG5" },*/
321 { "Big5-HKSCS", "BIG5-HKSCS" },
322 { "GB18030", "GB18030" },
323 /*{ "GBK", "GBK" },*/
324 { "ISO8859-1", "ISO-8859-1" },
325 { "ISO8859-11", "TIS-620" },
326 { "ISO8859-13", "ISO-8859-13" },
327 { "ISO8859-15", "ISO-8859-15" },
328 { "ISO8859-2", "ISO-8859-2" },
329 { "ISO8859-3", "ISO-8859-3" },
330 { "ISO8859-4", "ISO-8859-4" },
331 { "ISO8859-5", "ISO-8859-5" },
332 { "ISO8859-6", "ISO-8859-6" },
333 { "ISO8859-7", "ISO-8859-7" },
334 { "ISO8859-8", "ISO-8859-8" },
335 { "ISO8859-9", "ISO-8859-9" },
336 { "PCK", "SHIFT_JIS" },
337 { "TIS620.2533", "TIS-620" },
338 /*{ "UTF-8", "UTF-8" },*/
339 { "ansi-1251", "CP1251" },
340 { "cns11643", "EUC-TW" },
341 { "eucJP", "EUC-JP" },
342 { "gb2312", "GB2312" },
343 { "koi8-r", "KOI8-R" }
344 # define alias_table_defined
345 # endif
346 # if defined __minix /* Minix */
347 { "646", "ASCII" }
348 # define alias_table_defined
349 # endif
350 # if defined WINDOWS_NATIVE || defined __CYGWIN__ /* Windows */
351 { "CP1361", "JOHAB" },
352 { "CP20127", "ASCII" },
353 { "CP20866", "KOI8-R" },
354 { "CP20936", "GB2312" },
355 { "CP21866", "KOI8-RU" },
356 { "CP28591", "ISO-8859-1" },
357 { "CP28592", "ISO-8859-2" },
358 { "CP28593", "ISO-8859-3" },
359 { "CP28594", "ISO-8859-4" },
360 { "CP28595", "ISO-8859-5" },
361 { "CP28596", "ISO-8859-6" },
362 { "CP28597", "ISO-8859-7" },
363 { "CP28598", "ISO-8859-8" },
364 { "CP28599", "ISO-8859-9" },
365 { "CP28605", "ISO-8859-15" },
366 { "CP38598", "ISO-8859-8" },
367 { "CP51932", "EUC-JP" },
368 { "CP51936", "GB2312" },
369 { "CP51949", "EUC-KR" },
370 { "CP51950", "EUC-TW" },
371 { "CP54936", "GB18030" },
372 { "CP65001", "UTF-8" },
373 { "CP936", "GBK" }
374 # define alias_table_defined
375 # endif
376 # if defined OS2 /* OS/2 */
377 /* The list of encodings is taken from "List of OS/2 Codepages"
378 by Alex Taylor:
379 <http://altsan.org/os2/toolkits/uls/index.html#codepages>.
380 See also "IBM Globalization - Code page identifiers":
381 <https://www-01.ibm.com/software/globalization/cp/cp_cpgid.html>. */
382 { "CP1089", "ISO-8859-6" },
383 { "CP1208", "UTF-8" },
384 { "CP1381", "GB2312" },
385 { "CP1386", "GBK" },
386 { "CP3372", "EUC-JP" },
387 { "CP813", "ISO-8859-7" },
388 { "CP819", "ISO-8859-1" },
389 { "CP878", "KOI8-R" },
390 { "CP912", "ISO-8859-2" },
391 { "CP913", "ISO-8859-3" },
392 { "CP914", "ISO-8859-4" },
393 { "CP915", "ISO-8859-5" },
394 { "CP916", "ISO-8859-8" },
395 { "CP920", "ISO-8859-9" },
396 { "CP921", "ISO-8859-13" },
397 { "CP923", "ISO-8859-15" },
398 { "CP954", "EUC-JP" },
399 { "CP964", "EUC-TW" },
400 { "CP970", "EUC-KR" }
401 # define alias_table_defined
402 # endif
403 # if defined VMS /* OpenVMS */
404 /* The list of encodings is taken from the OpenVMS 7.3-1 documentation
405 "Compaq C Run-Time Library Reference Manual for OpenVMS systems"
406 section 10.7 "Handling Different Character Sets". */
407 { "DECHANYU", "DEC-HANYU" },
408 { "DECHANZI", "GB2312" },
409 { "DECKANJI", "DEC-KANJI" },
410 { "DECKOREAN", "EUC-KR" },
411 { "ISO8859-1", "ISO-8859-1" },
412 { "ISO8859-2", "ISO-8859-2" },
413 { "ISO8859-5", "ISO-8859-5" },
414 { "ISO8859-7", "ISO-8859-7" },
415 { "ISO8859-8", "ISO-8859-8" },
416 { "ISO8859-9", "ISO-8859-9" },
417 { "SDECKANJI", "EUC-JP" },
418 { "SJIS", "SHIFT_JIS" },
419 { "eucJP", "EUC-JP" },
420 { "eucTW", "EUC-TW" }
421 # define alias_table_defined
422 # endif
423 # ifndef alias_table_defined
424 /* Just a dummy entry, to avoid a C syntax error. */
425 { "", "" }
426 # endif
427 };
428
429 # endif
430
431 #else
432
433 /* On these platforms, we use a mapping from locale name to GNU canonical
434 encoding name. */
435
436 struct table_entry
437 {
438 const char locale[17+1];
439 const char canonical[11+1];
440 };
441
442 /* Table of platform-dependent mappings, sorted in ascending order. */
443 static const struct table_entry locale_table[] =
444 {
445 # if defined __FreeBSD__ /* FreeBSD 4.2 */
446 { "cs_CZ.ISO_8859-2", "ISO-8859-2" },
447 { "da_DK.DIS_8859-15", "ISO-8859-15" },
448 { "da_DK.ISO_8859-1", "ISO-8859-1" },
449 { "de_AT.DIS_8859-15", "ISO-8859-15" },
450 { "de_AT.ISO_8859-1", "ISO-8859-1" },
451 { "de_CH.DIS_8859-15", "ISO-8859-15" },
452 { "de_CH.ISO_8859-1", "ISO-8859-1" },
453 { "de_DE.DIS_8859-15", "ISO-8859-15" },
454 { "de_DE.ISO_8859-1", "ISO-8859-1" },
455 { "en_AU.DIS_8859-15", "ISO-8859-15" },
456 { "en_AU.ISO_8859-1", "ISO-8859-1" },
457 { "en_CA.DIS_8859-15", "ISO-8859-15" },
458 { "en_CA.ISO_8859-1", "ISO-8859-1" },
459 { "en_GB.DIS_8859-15", "ISO-8859-15" },
460 { "en_GB.ISO_8859-1", "ISO-8859-1" },
461 { "en_US.DIS_8859-15", "ISO-8859-15" },
462 { "en_US.ISO_8859-1", "ISO-8859-1" },
463 { "es_ES.DIS_8859-15", "ISO-8859-15" },
464 { "es_ES.ISO_8859-1", "ISO-8859-1" },
465 { "fi_FI.DIS_8859-15", "ISO-8859-15" },
466 { "fi_FI.ISO_8859-1", "ISO-8859-1" },
467 { "fr_BE.DIS_8859-15", "ISO-8859-15" },
468 { "fr_BE.ISO_8859-1", "ISO-8859-1" },
469 { "fr_CA.DIS_8859-15", "ISO-8859-15" },
470 { "fr_CA.ISO_8859-1", "ISO-8859-1" },
471 { "fr_CH.DIS_8859-15", "ISO-8859-15" },
472 { "fr_CH.ISO_8859-1", "ISO-8859-1" },
473 { "fr_FR.DIS_8859-15", "ISO-8859-15" },
474 { "fr_FR.ISO_8859-1", "ISO-8859-1" },
475 { "hr_HR.ISO_8859-2", "ISO-8859-2" },
476 { "hu_HU.ISO_8859-2", "ISO-8859-2" },
477 { "is_IS.DIS_8859-15", "ISO-8859-15" },
478 { "is_IS.ISO_8859-1", "ISO-8859-1" },
479 { "it_CH.DIS_8859-15", "ISO-8859-15" },
480 { "it_CH.ISO_8859-1", "ISO-8859-1" },
481 { "it_IT.DIS_8859-15", "ISO-8859-15" },
482 { "it_IT.ISO_8859-1", "ISO-8859-1" },
483 { "ja_JP.EUC", "EUC-JP" },
484 { "ja_JP.SJIS", "SHIFT_JIS" },
485 { "ja_JP.Shift_JIS", "SHIFT_JIS" },
486 { "ko_KR.EUC", "EUC-KR" },
487 { "la_LN.ASCII", "ASCII" },
488 { "la_LN.DIS_8859-15", "ISO-8859-15" },
489 { "la_LN.ISO_8859-1", "ISO-8859-1" },
490 { "la_LN.ISO_8859-2", "ISO-8859-2" },
491 { "la_LN.ISO_8859-4", "ISO-8859-4" },
492 { "lt_LN.ASCII", "ASCII" },
493 { "lt_LN.DIS_8859-15", "ISO-8859-15" },
494 { "lt_LN.ISO_8859-1", "ISO-8859-1" },
495 { "lt_LN.ISO_8859-2", "ISO-8859-2" },
496 { "lt_LT.ISO_8859-4", "ISO-8859-4" },
497 { "nl_BE.DIS_8859-15", "ISO-8859-15" },
498 { "nl_BE.ISO_8859-1", "ISO-8859-1" },
499 { "nl_NL.DIS_8859-15", "ISO-8859-15" },
500 { "nl_NL.ISO_8859-1", "ISO-8859-1" },
501 { "no_NO.DIS_8859-15", "ISO-8859-15" },
502 { "no_NO.ISO_8859-1", "ISO-8859-1" },
503 { "pl_PL.ISO_8859-2", "ISO-8859-2" },
504 { "pt_PT.DIS_8859-15", "ISO-8859-15" },
505 { "pt_PT.ISO_8859-1", "ISO-8859-1" },
506 { "ru_RU.CP866", "CP866" },
507 { "ru_RU.ISO_8859-5", "ISO-8859-5" },
508 { "ru_RU.KOI8-R", "KOI8-R" },
509 { "ru_SU.CP866", "CP866" },
510 { "ru_SU.ISO_8859-5", "ISO-8859-5" },
511 { "ru_SU.KOI8-R", "KOI8-R" },
512 { "sl_SI.ISO_8859-2", "ISO-8859-2" },
513 { "sv_SE.DIS_8859-15", "ISO-8859-15" },
514 { "sv_SE.ISO_8859-1", "ISO-8859-1" },
515 { "uk_UA.KOI8-U", "KOI8-U" },
516 { "zh_CN.EUC", "GB2312" },
517 { "zh_TW.BIG5", "BIG5" },
518 { "zh_TW.Big5", "BIG5" }
519 # define locale_table_defined
520 # endif
521 # if defined __DJGPP__ /* DOS / DJGPP 2.03 */
522 /* The encodings given here may not all be correct.
523 If you find that the encoding given for your language and
524 country is not the one your DOS machine actually uses, just
525 correct it in this file, and send a mail to
526 Juan Manuel Guerrero <juan.guerrero@gmx.de>
527 and <bug-gnulib@gnu.org>. */
528 { "C", "ASCII" },
529 { "ar", "CP864" },
530 { "ar_AE", "CP864" },
531 { "ar_DZ", "CP864" },
532 { "ar_EG", "CP864" },
533 { "ar_IQ", "CP864" },
534 { "ar_IR", "CP864" },
535 { "ar_JO", "CP864" },
536 { "ar_KW", "CP864" },
537 { "ar_MA", "CP864" },
538 { "ar_OM", "CP864" },
539 { "ar_QA", "CP864" },
540 { "ar_SA", "CP864" },
541 { "ar_SY", "CP864" },
542 { "be", "CP866" },
543 { "be_BE", "CP866" },
544 { "bg", "CP866" }, /* not CP855 ?? */
545 { "bg_BG", "CP866" }, /* not CP855 ?? */
546 { "ca", "CP850" },
547 { "ca_ES", "CP850" },
548 { "cs", "CP852" },
549 { "cs_CZ", "CP852" },
550 { "da", "CP865" }, /* not CP850 ?? */
551 { "da_DK", "CP865" }, /* not CP850 ?? */
552 { "de", "CP850" },
553 { "de_AT", "CP850" },
554 { "de_CH", "CP850" },
555 { "de_DE", "CP850" },
556 { "el", "CP869" },
557 { "el_GR", "CP869" },
558 { "en", "CP850" },
559 { "en_AU", "CP850" }, /* not CP437 ?? */
560 { "en_CA", "CP850" },
561 { "en_GB", "CP850" },
562 { "en_NZ", "CP437" },
563 { "en_US", "CP437" },
564 { "en_ZA", "CP850" }, /* not CP437 ?? */
565 { "eo", "CP850" },
566 { "eo_EO", "CP850" },
567 { "es", "CP850" },
568 { "es_AR", "CP850" },
569 { "es_BO", "CP850" },
570 { "es_CL", "CP850" },
571 { "es_CO", "CP850" },
572 { "es_CR", "CP850" },
573 { "es_CU", "CP850" },
574 { "es_DO", "CP850" },
575 { "es_EC", "CP850" },
576 { "es_ES", "CP850" },
577 { "es_GT", "CP850" },
578 { "es_HN", "CP850" },
579 { "es_MX", "CP850" },
580 { "es_NI", "CP850" },
581 { "es_PA", "CP850" },
582 { "es_PE", "CP850" },
583 { "es_PY", "CP850" },
584 { "es_SV", "CP850" },
585 { "es_UY", "CP850" },
586 { "es_VE", "CP850" },
587 { "et", "CP850" },
588 { "et_EE", "CP850" },
589 { "eu", "CP850" },
590 { "eu_ES", "CP850" },
591 { "fi", "CP850" },
592 { "fi_FI", "CP850" },
593 { "fr", "CP850" },
594 { "fr_BE", "CP850" },
595 { "fr_CA", "CP850" },
596 { "fr_CH", "CP850" },
597 { "fr_FR", "CP850" },
598 { "ga", "CP850" },
599 { "ga_IE", "CP850" },
600 { "gd", "CP850" },
601 { "gd_GB", "CP850" },
602 { "gl", "CP850" },
603 { "gl_ES", "CP850" },
604 { "he", "CP862" },
605 { "he_IL", "CP862" },
606 { "hr", "CP852" },
607 { "hr_HR", "CP852" },
608 { "hu", "CP852" },
609 { "hu_HU", "CP852" },
610 { "id", "CP850" }, /* not CP437 ?? */
611 { "id_ID", "CP850" }, /* not CP437 ?? */
612 { "is", "CP861" }, /* not CP850 ?? */
613 { "is_IS", "CP861" }, /* not CP850 ?? */
614 { "it", "CP850" },
615 { "it_CH", "CP850" },
616 { "it_IT", "CP850" },
617 { "ja", "CP932" },
618 { "ja_JP", "CP932" },
619 { "kr", "CP949" }, /* not CP934 ?? */
620 { "kr_KR", "CP949" }, /* not CP934 ?? */
621 { "lt", "CP775" },
622 { "lt_LT", "CP775" },
623 { "lv", "CP775" },
624 { "lv_LV", "CP775" },
625 { "mk", "CP866" }, /* not CP855 ?? */
626 { "mk_MK", "CP866" }, /* not CP855 ?? */
627 { "mt", "CP850" },
628 { "mt_MT", "CP850" },
629 { "nb", "CP865" }, /* not CP850 ?? */
630 { "nb_NO", "CP865" }, /* not CP850 ?? */
631 { "nl", "CP850" },
632 { "nl_BE", "CP850" },
633 { "nl_NL", "CP850" },
634 { "nn", "CP865" }, /* not CP850 ?? */
635 { "nn_NO", "CP865" }, /* not CP850 ?? */
636 { "no", "CP865" }, /* not CP850 ?? */
637 { "no_NO", "CP865" }, /* not CP850 ?? */
638 { "pl", "CP852" },
639 { "pl_PL", "CP852" },
640 { "pt", "CP850" },
641 { "pt_BR", "CP850" },
642 { "pt_PT", "CP850" },
643 { "ro", "CP852" },
644 { "ro_RO", "CP852" },
645 { "ru", "CP866" },
646 { "ru_RU", "CP866" },
647 { "sk", "CP852" },
648 { "sk_SK", "CP852" },
649 { "sl", "CP852" },
650 { "sl_SI", "CP852" },
651 { "sq", "CP852" },
652 { "sq_AL", "CP852" },
653 { "sr", "CP852" }, /* CP852 or CP866 or CP855 ?? */
654 { "sr_CS", "CP852" }, /* CP852 or CP866 or CP855 ?? */
655 { "sr_YU", "CP852" }, /* CP852 or CP866 or CP855 ?? */
656 { "sv", "CP850" },
657 { "sv_SE", "CP850" },
658 { "th", "CP874" },
659 { "th_TH", "CP874" },
660 { "tr", "CP857" },
661 { "tr_TR", "CP857" },
662 { "uk", "CP1125" },
663 { "uk_UA", "CP1125" },
664 { "zh_CN", "GBK" },
665 { "zh_TW", "CP950" } /* not CP938 ?? */
666 # define locale_table_defined
667 # endif
668 # ifndef locale_table_defined
669 /* Just a dummy entry, to avoid a C syntax error. */
670 { "", "" }
671 # endif
672 };
673
674 #endif
675
676
677 /* Determine the current locale's character encoding, and canonicalize it
678 into one of the canonical names listed in localcharset.h.
679 The result must not be freed; it is statically allocated.
680 If the canonical name cannot be determined, the result is a non-canonical
681 name. */
682
683 #ifdef STATIC
684 STATIC
685 #endif
686 const char *
locale_charset(void)687 locale_charset (void)
688 {
689 const char *codeset;
690
691 #if HAVE_LANGINFO_CODESET || defined WINDOWS_NATIVE || defined OS2
692
693 # if HAVE_LANGINFO_CODESET
694
695 /* Most systems support nl_langinfo (CODESET) nowadays. */
696 codeset = nl_langinfo (CODESET);
697
698 # ifdef __CYGWIN__
699 /* Cygwin < 1.7 does not have locales. nl_langinfo (CODESET) always
700 returns "US-ASCII". Return the suffix of the locale name from the
701 environment variables (if present) or the codepage as a number. */
702 if (codeset != NULL && strcmp (codeset, "US-ASCII") == 0)
703 {
704 const char *locale;
705 static char buf[2 + 10 + 1];
706
707 locale = getenv ("LC_ALL");
708 if (locale == NULL || locale[0] == '\0')
709 {
710 locale = getenv ("LC_CTYPE");
711 if (locale == NULL || locale[0] == '\0')
712 locale = getenv ("LANG");
713 }
714 if (locale != NULL && locale[0] != '\0')
715 {
716 /* If the locale name contains an encoding after the dot, return
717 it. */
718 const char *dot = strchr (locale, '.');
719
720 if (dot != NULL)
721 {
722 const char *modifier;
723
724 dot++;
725 /* Look for the possible @... trailer and remove it, if any. */
726 modifier = strchr (dot, '@');
727 if (modifier == NULL)
728 return dot;
729 if (modifier - dot < sizeof (buf))
730 {
731 memcpy (buf, dot, modifier - dot);
732 buf [modifier - dot] = '\0';
733 return buf;
734 }
735 }
736 }
737
738 /* The Windows API has a function returning the locale's codepage as a
739 number: GetACP(). This encoding is used by Cygwin, unless the user
740 has set the environment variable CYGWIN=codepage:oem (which very few
741 people do).
742 Output directed to console windows needs to be converted (to
743 GetOEMCP() if the console is using a raster font, or to
744 GetConsoleOutputCP() if it is using a TrueType font). Cygwin does
745 this conversion transparently (see winsup/cygwin/fhandler_console.cc),
746 converting to GetConsoleOutputCP(). This leads to correct results,
747 except when SetConsoleOutputCP has been called and a raster font is
748 in use. */
749 sprintf (buf, "CP%u", GetACP ());
750 codeset = buf;
751 }
752 # endif
753
754 if (codeset == NULL)
755 /* The canonical name cannot be determined. */
756 codeset = "";
757
758 # elif defined WINDOWS_NATIVE
759
760 static char buf[2 + 10 + 1];
761
762 /* The Windows API has a function returning the locale's codepage as
763 a number, but the value doesn't change according to what the
764 'setlocale' call specified. So we use it as a last resort, in
765 case the string returned by 'setlocale' doesn't specify the
766 codepage. */
767 char *current_locale = setlocale (LC_ALL, NULL);
768 char *pdot;
769
770 /* If they set different locales for different categories,
771 'setlocale' will return a semi-colon separated list of locale
772 values. To make sure we use the correct one, we choose LC_CTYPE. */
773 if (strchr (current_locale, ';'))
774 current_locale = setlocale (LC_CTYPE, NULL);
775
776 pdot = strrchr (current_locale, '.');
777 if (pdot && 2 + strlen (pdot + 1) + 1 <= sizeof (buf))
778 sprintf (buf, "CP%s", pdot + 1);
779 else
780 {
781 /* The Windows API has a function returning the locale's codepage as a
782 number: GetACP().
783 When the output goes to a console window, it needs to be provided in
784 GetOEMCP() encoding if the console is using a raster font, or in
785 GetConsoleOutputCP() encoding if it is using a TrueType font.
786 But in GUI programs and for output sent to files and pipes, GetACP()
787 encoding is the best bet. */
788 sprintf (buf, "CP%u", GetACP ());
789 }
790 codeset = buf;
791
792 # elif defined OS2
793
794 const char *locale;
795 static char buf[2 + 10 + 1];
796 ULONG cp[3];
797 ULONG cplen;
798
799 codeset = NULL;
800
801 /* Allow user to override the codeset, as set in the operating system,
802 with standard language environment variables. */
803 locale = getenv ("LC_ALL");
804 if (locale == NULL || locale[0] == '\0')
805 {
806 locale = getenv ("LC_CTYPE");
807 if (locale == NULL || locale[0] == '\0')
808 locale = getenv ("LANG");
809 }
810 if (locale != NULL && locale[0] != '\0')
811 {
812 /* If the locale name contains an encoding after the dot, return it. */
813 const char *dot = strchr (locale, '.');
814
815 if (dot != NULL)
816 {
817 const char *modifier;
818
819 dot++;
820 /* Look for the possible @... trailer and remove it, if any. */
821 modifier = strchr (dot, '@');
822 if (modifier == NULL)
823 return dot;
824 if (modifier - dot < sizeof (buf))
825 {
826 memcpy (buf, dot, modifier - dot);
827 buf [modifier - dot] = '\0';
828 return buf;
829 }
830 }
831
832 /* For the POSIX locale, don't use the system's codepage. */
833 if (strcmp (locale, "C") == 0 || strcmp (locale, "POSIX") == 0)
834 codeset = "";
835 }
836
837 if (codeset == NULL)
838 {
839 /* OS/2 has a function returning the locale's codepage as a number. */
840 if (DosQueryCp (sizeof (cp), cp, &cplen))
841 codeset = "";
842 else
843 {
844 sprintf (buf, "CP%u", cp[0]);
845 codeset = buf;
846 }
847 }
848
849 # else
850
851 # error "Add code for other platforms here."
852
853 # endif
854
855 /* Resolve alias. */
856 {
857 # ifdef alias_table_defined
858 /* On some platforms, UTF-8 locales are the most frequently used ones.
859 Speed up the common case and slow down the less common cases by
860 testing for this case first. */
861 # if defined __OpenBSD__ || (defined __APPLE__ && defined __MACH__) || defined __sun || defined __CYGWIN__
862 if (strcmp (codeset, "UTF-8") == 0)
863 goto done_table_lookup;
864 else
865 # endif
866 {
867 const struct table_entry * const table = alias_table;
868 size_t const table_size =
869 sizeof (alias_table) / sizeof (struct table_entry);
870 /* The table is sorted. Perform a binary search. */
871 size_t hi = table_size;
872 size_t lo = 0;
873 while (lo < hi)
874 {
875 /* Invariant:
876 for i < lo, strcmp (table[i].alias, codeset) < 0,
877 for i >= hi, strcmp (table[i].alias, codeset) > 0. */
878 size_t mid = (hi + lo) >> 1; /* >= lo, < hi */
879 int cmp = strcmp (table[mid].alias, codeset);
880 if (cmp < 0)
881 lo = mid + 1;
882 else if (cmp > 0)
883 hi = mid;
884 else
885 {
886 /* Found an i with
887 strcmp (table[i].alias, codeset) == 0. */
888 codeset = table[mid].canonical;
889 goto done_table_lookup;
890 }
891 }
892 }
893 if (0)
894 done_table_lookup: ;
895 else
896 # endif
897 {
898 /* Did not find it in the table. */
899 /* On Mac OS X, all modern locales use the UTF-8 encoding.
900 BeOS and Haiku have a single locale, and it has UTF-8 encoding. */
901 # if (defined __APPLE__ && defined __MACH__) || defined __BEOS__ || defined __HAIKU__
902 codeset = "UTF-8";
903 # else
904 /* Don't return an empty string. GNU libc and GNU libiconv interpret
905 the empty string as denoting "the locale's character encoding",
906 thus GNU libiconv would call this function a second time. */
907 if (codeset[0] == '\0')
908 codeset = "ASCII";
909 # endif
910 }
911 }
912
913 #else
914
915 /* On old systems which lack it, use setlocale or getenv. */
916 const char *locale = NULL;
917
918 /* But most old systems don't have a complete set of locales. Some
919 (like DJGPP) have only the C locale. Therefore we don't use setlocale
920 here; it would return "C" when it doesn't support the locale name the
921 user has set. */
922 # if 0
923 locale = setlocale (LC_CTYPE, NULL);
924 # endif
925 if (locale == NULL || locale[0] == '\0')
926 {
927 locale = getenv ("LC_ALL");
928 if (locale == NULL || locale[0] == '\0')
929 {
930 locale = getenv ("LC_CTYPE");
931 if (locale == NULL || locale[0] == '\0')
932 locale = getenv ("LANG");
933 if (locale == NULL)
934 locale = "";
935 }
936 }
937
938 /* Map locale name to canonical encoding name. */
939 {
940 # ifdef locale_table_defined
941 const struct table_entry * const table = locale_table;
942 size_t const table_size =
943 sizeof (locale_table) / sizeof (struct table_entry);
944 /* The table is sorted. Perform a binary search. */
945 size_t hi = table_size;
946 size_t lo = 0;
947 while (lo < hi)
948 {
949 /* Invariant:
950 for i < lo, strcmp (table[i].locale, locale) < 0,
951 for i >= hi, strcmp (table[i].locale, locale) > 0. */
952 size_t mid = (hi + lo) >> 1; /* >= lo, < hi */
953 int cmp = strcmp (table[mid].locale, locale);
954 if (cmp < 0)
955 lo = mid + 1;
956 else if (cmp > 0)
957 hi = mid;
958 else
959 {
960 /* Found an i with
961 strcmp (table[i].locale, locale) == 0. */
962 codeset = table[mid].canonical;
963 goto done_table_lookup;
964 }
965 }
966 if (0)
967 done_table_lookup: ;
968 else
969 # endif
970 {
971 /* Did not find it in the table. */
972 /* On Mac OS X, all modern locales use the UTF-8 encoding.
973 BeOS and Haiku have a single locale, and it has UTF-8 encoding. */
974 # if (defined __APPLE__ && defined __MACH__) || defined __BEOS__ || defined __HAIKU__
975 codeset = "UTF-8";
976 # else
977 /* The canonical name cannot be determined. */
978 /* Don't return an empty string. GNU libc and GNU libiconv interpret
979 the empty string as denoting "the locale's character encoding",
980 thus GNU libiconv would call this function a second time. */
981 codeset = "ASCII";
982 # endif
983 }
984 }
985
986 #endif
987
988 #ifdef DARWIN7
989 /* Mac OS X sets MB_CUR_MAX to 1 when LC_ALL=C, and "UTF-8"
990 (the default codeset) does not work when MB_CUR_MAX is 1. */
991 if (strcmp (codeset, "UTF-8") == 0 && MB_CUR_MAX_L (uselocale (NULL)) <= 1)
992 codeset = "ASCII";
993 #endif
994
995 return codeset;
996 }
997