1 /* Load needed message catalogs.
2 Copyright (C) 1995-1999, 2000, 2001 Free Software Foundation, Inc.
3
4 This program is free software; you can redistribute it and/or modify it
5 under the terms of the GNU Library General Public License as published
6 by the Free Software Foundation; either version 2, or (at your option)
7 any later version.
8
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 Library General Public License for more details.
13
14 You should have received a copy of the GNU Library General Public
15 License along with this program; if not, write to the Free Software
16 Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307,
17 USA. */
18
19 /* Tell glibc's <string.h> to provide a prototype for mempcpy().
20 This must come before <config.h> because <config.h> may include
21 <features.h>, and once <features.h> has been included, it's too late. */
22 #ifndef _GNU_SOURCE
23 # define _GNU_SOURCE 1
24 #endif
25
26 #ifdef HAVE_CONFIG_H
27 # include <config.h>
28 #endif
29
30 #include <ctype.h>
31 #include <errno.h>
32 #include <fcntl.h>
33 #include <sys/types.h>
34 #include <sys/stat.h>
35
36 #ifdef __GNUC__
37 # define alloca __builtin_alloca
38 # define HAVE_ALLOCA 1
39 #else
40 # if defined HAVE_ALLOCA_H || defined _LIBC
41 # include <alloca.h>
42 # else
43 # ifdef _AIX
44 #pragma alloca
45 # else
46 # ifndef alloca
47 char *alloca ();
48 # endif
49 # endif
50 # endif
51 #endif
52
53 #include <stdlib.h>
54 #include <string.h>
55
56 #if defined HAVE_UNISTD_H || defined _LIBC
57 # include <unistd.h>
58 #endif
59
60 #ifdef _LIBC
61 # include <langinfo.h>
62 # include <locale.h>
63 #endif
64
65 #if (defined HAVE_MMAP && defined HAVE_MUNMAP && !defined DISALLOW_MMAP) \
66 || (defined _LIBC && defined _POSIX_MAPPED_FILES)
67 # include <sys/mman.h>
68 # undef HAVE_MMAP
69 # define HAVE_MMAP 1
70 #else
71 # undef HAVE_MMAP
72 #endif
73
74 #include "gettext.h"
75 #include "gettextP.h"
76
77 #ifdef _LIBC
78 # include "../locale/localeinfo.h"
79 #endif
80
81 /* @@ end of prolog @@ */
82
83 #ifdef _LIBC
84 /* Rename the non ISO C functions. This is required by the standard
85 because some ISO C functions will require linking with this object
86 file and the name space must not be polluted. */
87 # define open __open
88 # define close __close
89 # define read __read
90 # define mmap __mmap
91 # define munmap __munmap
92 #endif
93
94 /* Names for the libintl functions are a problem. They must not clash
95 with existing names and they should follow ANSI C. But this source
96 code is also used in GNU C Library where the names have a __
97 prefix. So we have to make a difference here. */
98 #ifdef _LIBC
99 # define PLURAL_PARSE __gettextparse
100 #else
101 # define PLURAL_PARSE gettextparse__
102 #endif
103
104 /* For those losing systems which don't have `alloca' we have to add
105 some additional code emulating it. */
106 #ifdef HAVE_ALLOCA
107 # define freea(p) /* nothing */
108 #else
109 # define alloca(n) malloc (n)
110 # define freea(p) free (p)
111 #endif
112
113 /* For systems that distinguish between text and binary I/O.
114 O_BINARY is usually declared in <fcntl.h>. */
115 #if !defined O_BINARY && defined _O_BINARY
116 /* For MSC-compatible compilers. */
117 # define O_BINARY _O_BINARY
118 # define O_TEXT _O_TEXT
119 #endif
120 #ifdef __BEOS__
121 /* BeOS 5 has O_BINARY and O_TEXT, but they have no effect. */
122 # undef O_BINARY
123 # undef O_TEXT
124 #endif
125 /* On reasonable systems, binary I/O is the default. */
126 #ifndef O_BINARY
127 # define O_BINARY 0
128 #endif
129
130 /* We need a sign, whether a new catalog was loaded, which can be associated
131 with all translations. This is important if the translations are
132 cached by one of GCC's features. */
133 int _nl_msg_cat_cntr;
134
135 #if (defined __GNUC__ && !defined __APPLE_CC__) \
136 || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
137
138 /* These structs are the constant expression for the germanic plural
139 form determination. It represents the expression "n != 1". */
140 static const struct expression plvar =
141 {
142 .nargs = 0,
143 .operation = var,
144 };
145 static const struct expression plone =
146 {
147 .nargs = 0,
148 .operation = num,
149 .val =
150 {
151 .num = 1
152 }
153 };
154 static struct expression germanic_plural =
155 {
156 .nargs = 2,
157 .operation = not_equal,
158 .val =
159 {
160 .args =
161 {
162 [0] = (struct expression *) &plvar,
163 [1] = (struct expression *) &plone
164 }
165 }
166 };
167
168 # define INIT_GERMANIC_PLURAL()
169
170 #else
171
172 /* For compilers without support for ISO C 99 struct/union initializers:
173 Initialization at run-time. */
174
175 static struct expression plvar;
176 static struct expression plone;
177 static struct expression germanic_plural;
178
179 static void
init_germanic_plural()180 init_germanic_plural ()
181 {
182 if (plone.val.num == 0)
183 {
184 plvar.nargs = 0;
185 plvar.operation = var;
186
187 plone.nargs = 0;
188 plone.operation = num;
189 plone.val.num = 1;
190
191 germanic_plural.nargs = 2;
192 germanic_plural.operation = not_equal;
193 germanic_plural.val.args[0] = &plvar;
194 germanic_plural.val.args[1] = &plone;
195 }
196 }
197
198 # define INIT_GERMANIC_PLURAL() init_germanic_plural ()
199
200 #endif
201
202
203 /* Initialize the codeset dependent parts of an opened message catalog.
204 Return the header entry. */
205 const char *
206 internal_function
_nl_init_domain_conv(domain_file,domain,domainbinding)207 _nl_init_domain_conv (domain_file, domain, domainbinding)
208 struct loaded_l10nfile *domain_file;
209 struct loaded_domain *domain;
210 struct binding *domainbinding;
211 {
212 /* Find out about the character set the file is encoded with.
213 This can be found (in textual form) in the entry "". If this
214 entry does not exist or if this does not contain the `charset='
215 information, we will assume the charset matches the one the
216 current locale and we don't have to perform any conversion. */
217 char *nullentry;
218 size_t nullentrylen;
219
220 /* Preinitialize fields, to avoid recursion during _nl_find_msg. */
221 domain->codeset_cntr =
222 (domainbinding != NULL ? domainbinding->codeset_cntr : 0);
223 #ifdef _LIBC
224 domain->conv = (__gconv_t) -1;
225 #else
226 # if HAVE_ICONV
227 domain->conv = (iconv_t) -1;
228 # endif
229 #endif
230 domain->conv_tab = NULL;
231
232 /* Get the header entry. */
233 nullentry = _nl_find_msg (domain_file, domainbinding, "", &nullentrylen);
234
235 if (nullentry != NULL)
236 {
237 #if defined _LIBC || HAVE_ICONV
238 const char *charsetstr;
239
240 charsetstr = strstr (nullentry, "charset=");
241 if (charsetstr != NULL)
242 {
243 size_t len;
244 char *charset;
245 const char *outcharset;
246
247 charsetstr += strlen ("charset=");
248 len = strcspn (charsetstr, " \t\n");
249
250 charset = (char *) alloca (len + 1);
251 # if defined _LIBC || HAVE_MEMPCPY
252 *((char *) mempcpy (charset, charsetstr, len)) = '\0';
253 # else
254 memcpy (charset, charsetstr, len);
255 charset[len] = '\0';
256 # endif
257
258 /* The output charset should normally be determined by the
259 locale. But sometimes the locale is not used or not correctly
260 set up, so we provide a possibility for the user to override
261 this. Moreover, the value specified through
262 bind_textdomain_codeset overrides both. */
263 if (domainbinding != NULL && domainbinding->codeset != NULL)
264 outcharset = domainbinding->codeset;
265 else
266 {
267 outcharset = getenv ("OUTPUT_CHARSET");
268 if (outcharset == NULL || outcharset[0] == '\0')
269 {
270 # ifdef _LIBC
271 outcharset = (*_nl_current[LC_CTYPE])->values[_NL_ITEM_INDEX (CODESET)].string;
272 # else
273 # if HAVE_ICONV
274 extern const char *locale_charset (void);
275 outcharset = locale_charset ();
276 # endif
277 # endif
278 }
279 }
280
281 # ifdef _LIBC
282 /* We always want to use transliteration. */
283 outcharset = norm_add_slashes (outcharset, "TRANSLIT");
284 charset = norm_add_slashes (charset, NULL);
285 if (__gconv_open (outcharset, charset, &domain->conv,
286 GCONV_AVOID_NOCONV)
287 != __GCONV_OK)
288 domain->conv = (__gconv_t) -1;
289 # else
290 # if HAVE_ICONV
291 /* When using GNU libiconv, we want to use transliteration. */
292 # if _LIBICONV_VERSION >= 0x0105
293 len = strlen (outcharset);
294 {
295 char *tmp = (char *) alloca (len + 10 + 1);
296 memcpy (tmp, outcharset, len);
297 memcpy (tmp + len, "//TRANSLIT", 10 + 1);
298 outcharset = tmp;
299 }
300 # endif
301 domain->conv = iconv_open (outcharset, charset);
302 # if _LIBICONV_VERSION >= 0x0105
303 freea (outcharset);
304 # endif
305 # endif
306 # endif
307
308 freea (charset);
309 }
310 #endif /* _LIBC || HAVE_ICONV */
311 }
312
313 return nullentry;
314 }
315
316 /* Frees the codeset dependent parts of an opened message catalog. */
317 void
318 internal_function
_nl_free_domain_conv(domain)319 _nl_free_domain_conv (domain)
320 struct loaded_domain *domain;
321 {
322 if (domain->conv_tab != NULL && domain->conv_tab != (char **) -1)
323 free (domain->conv_tab);
324
325 #ifdef _LIBC
326 if (domain->conv != (__gconv_t) -1)
327 __gconv_close (domain->conv);
328 #else
329 # if HAVE_ICONV
330 if (domain->conv != (iconv_t) -1)
331 iconv_close (domain->conv);
332 # endif
333 #endif
334 }
335
336 /* Load the message catalogs specified by FILENAME. If it is no valid
337 message catalog do nothing. */
338 void
339 internal_function
_nl_load_domain(domain_file,domainbinding)340 _nl_load_domain (domain_file, domainbinding)
341 struct loaded_l10nfile *domain_file;
342 struct binding *domainbinding;
343 {
344 int fd;
345 size_t size;
346 #ifdef _LIBC
347 struct stat64 st;
348 #else
349 struct stat st;
350 #endif
351 struct mo_file_header *data = (struct mo_file_header *) -1;
352 int use_mmap = 0;
353 struct loaded_domain *domain;
354 const char *nullentry;
355
356 domain_file->decided = 1;
357 domain_file->data = NULL;
358
359 /* Note that it would be useless to store domainbinding in domain_file
360 because domainbinding might be == NULL now but != NULL later (after
361 a call to bind_textdomain_codeset). */
362
363 /* If the record does not represent a valid locale the FILENAME
364 might be NULL. This can happen when according to the given
365 specification the locale file name is different for XPG and CEN
366 syntax. */
367 if (domain_file->filename == NULL)
368 return;
369
370 /* Try to open the addressed file. */
371 fd = open (domain_file->filename, O_RDONLY | O_BINARY);
372 if (fd == -1)
373 return;
374
375 /* We must know about the size of the file. */
376 if (
377 #ifdef _LIBC
378 __builtin_expect (fstat64 (fd, &st) != 0, 0)
379 #else
380 __builtin_expect (fstat (fd, &st) != 0, 0)
381 #endif
382 || __builtin_expect ((size = (size_t) st.st_size) != st.st_size, 0)
383 || __builtin_expect (size < sizeof (struct mo_file_header), 0))
384 {
385 /* Something went wrong. */
386 close (fd);
387 return;
388 }
389
390 #ifdef HAVE_MMAP
391 /* Now we are ready to load the file. If mmap() is available we try
392 this first. If not available or it failed we try to load it. */
393 data = (struct mo_file_header *) mmap (NULL, size, PROT_READ,
394 MAP_PRIVATE, fd, 0);
395
396 if (__builtin_expect (data != (struct mo_file_header *) -1, 1))
397 {
398 /* mmap() call was successful. */
399 close (fd);
400 use_mmap = 1;
401 }
402 #endif
403
404 /* If the data is not yet available (i.e. mmap'ed) we try to load
405 it manually. */
406 if (data == (struct mo_file_header *) -1)
407 {
408 size_t to_read;
409 char *read_ptr;
410
411 data = (struct mo_file_header *) malloc (size);
412 if (data == NULL)
413 return;
414
415 to_read = size;
416 read_ptr = (char *) data;
417 do
418 {
419 long int nb = (long int) read (fd, read_ptr, to_read);
420 if (nb <= 0)
421 {
422 #ifdef EINTR
423 if (nb == -1 && errno == EINTR)
424 continue;
425 #endif
426 close (fd);
427 return;
428 }
429 read_ptr += nb;
430 to_read -= nb;
431 }
432 while (to_read > 0);
433
434 close (fd);
435 }
436
437 /* Using the magic number we can test whether it really is a message
438 catalog file. */
439 if (__builtin_expect (data->magic != _MAGIC && data->magic != _MAGIC_SWAPPED,
440 0))
441 {
442 /* The magic number is wrong: not a message catalog file. */
443 #ifdef HAVE_MMAP
444 if (use_mmap)
445 munmap ((caddr_t) data, size);
446 else
447 #endif
448 free (data);
449 return;
450 }
451
452 domain = (struct loaded_domain *) malloc (sizeof (struct loaded_domain));
453 if (domain == NULL)
454 return;
455 domain_file->data = domain;
456
457 domain->data = (char *) data;
458 domain->use_mmap = use_mmap;
459 domain->mmap_size = size;
460 domain->must_swap = data->magic != _MAGIC;
461
462 /* Fill in the information about the available tables. */
463 switch (W (domain->must_swap, data->revision))
464 {
465 case 0:
466 domain->nstrings = W (domain->must_swap, data->nstrings);
467 domain->orig_tab = (struct string_desc *)
468 ((char *) data + W (domain->must_swap, data->orig_tab_offset));
469 domain->trans_tab = (struct string_desc *)
470 ((char *) data + W (domain->must_swap, data->trans_tab_offset));
471 domain->hash_size = W (domain->must_swap, data->hash_tab_size);
472 domain->hash_tab = (nls_uint32 *)
473 ((char *) data + W (domain->must_swap, data->hash_tab_offset));
474 break;
475 default:
476 /* This is an invalid revision. */
477 #ifdef HAVE_MMAP
478 if (use_mmap)
479 munmap ((caddr_t) data, size);
480 else
481 #endif
482 free (data);
483 free (domain);
484 domain_file->data = NULL;
485 return;
486 }
487
488 /* Now initialize the character set converter from the character set
489 the file is encoded with (found in the header entry) to the domain's
490 specified character set or the locale's character set. */
491 nullentry = _nl_init_domain_conv (domain_file, domain, domainbinding);
492
493 /* Also look for a plural specification. */
494 if (nullentry != NULL)
495 {
496 const char *plural;
497 const char *nplurals;
498
499 plural = strstr (nullentry, "plural=");
500 nplurals = strstr (nullentry, "nplurals=");
501 if (plural == NULL || nplurals == NULL)
502 goto no_plural;
503 else
504 {
505 /* First get the number. */
506 char *endp;
507 unsigned long int n;
508 struct parse_args args;
509
510 nplurals += 9;
511 while (*nplurals != '\0' && isspace (*nplurals))
512 ++nplurals;
513 #if defined HAVE_STRTOUL || defined _LIBC
514 n = strtoul (nplurals, &endp, 10);
515 #else
516 for (endp = nplurals, n = 0; *endp >= '0' && *endp <= '9'; endp++)
517 n = n * 10 + (*endp - '0');
518 #endif
519 domain->nplurals = n;
520 if (nplurals == endp)
521 goto no_plural;
522
523 /* Due to the restrictions bison imposes onto the interface of the
524 scanner function we have to put the input string and the result
525 passed up from the parser into the same structure which address
526 is passed down to the parser. */
527 plural += 7;
528 args.cp = plural;
529 if (PLURAL_PARSE (&args) != 0)
530 goto no_plural;
531 domain->plural = args.res;
532 }
533 }
534 else
535 {
536 /* By default we are using the Germanic form: singular form only
537 for `one', the plural form otherwise. Yes, this is also what
538 English is using since English is a Germanic language. */
539 no_plural:
540 INIT_GERMANIC_PLURAL ();
541 domain->plural = &germanic_plural;
542 domain->nplurals = 2;
543 }
544 }
545
546
547 #ifdef _LIBC
548 void
549 internal_function
_nl_unload_domain(domain)550 _nl_unload_domain (domain)
551 struct loaded_domain *domain;
552 {
553 if (domain->plural != &germanic_plural)
554 __gettext_free_exp (domain->plural);
555
556 _nl_free_domain_conv (domain);
557
558 # ifdef _POSIX_MAPPED_FILES
559 if (domain->use_mmap)
560 munmap ((caddr_t) domain->data, domain->mmap_size);
561 else
562 # endif /* _POSIX_MAPPED_FILES */
563 free ((void *) domain->data);
564
565 free (domain);
566 }
567 #endif
568