Update.
[glibc.git] / intl / dcgettext.c
blobd482da39b8469ac21fb211b5f2c11188be4eedb5
1 /* Implementation of the dcgettext(3) function.
2 Copyright (C) 1995, 1996, 1997, 1998, 1999 Free Software Foundation, Inc.
4 This file is part of the GNU C Library. Its master source is NOT part of
5 the C library, however.
7 The GNU C Library is free software; you can redistribute it and/or
8 modify it under the terms of the GNU Library General Public License as
9 published by the Free Software Foundation; either version 2 of the
10 License, or (at your option) any later version.
12 The GNU C Library is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 Library General Public License for more details.
17 You should have received a copy of the GNU Library General Public
18 License along with the GNU C Library; see the file COPYING.LIB. If not,
19 write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
20 Boston, MA 02111-1307, USA. */
22 #ifdef HAVE_CONFIG_H
23 # include <config.h>
24 #endif
26 #include <sys/types.h>
28 #if defined __GNUC__ && !defined C_ALLOCA
29 # define alloca __builtin_alloca
30 # define HAVE_ALLOCA 1
31 #else
32 # if (defined HAVE_ALLOCA_H || defined _LIBC) && !defined C_ALLOCA
33 # include <alloca.h>
34 # else
35 # ifdef _AIX
36 #pragma alloca
37 # else
38 # ifndef alloca
39 char *alloca ();
40 # endif
41 # endif
42 # endif
43 #endif
45 #include <errno.h>
46 #ifndef errno
47 extern int errno;
48 #endif
49 #ifndef __set_errno
50 # define __set_errno(val) errno = (val)
51 #endif
53 #if defined STDC_HEADERS || defined _LIBC
54 # include <stdlib.h>
55 #else
56 char *getenv ();
57 # ifdef HAVE_MALLOC_H
58 # include <malloc.h>
59 # else
60 void free ();
61 # endif
62 #endif
64 #if defined HAVE_STRING_H || defined _LIBC
65 # ifndef _GNU_SOURCE
66 # define _GNU_SOURCE 1
67 # endif
68 # include <string.h>
69 #else
70 # include <strings.h>
71 #endif
72 #if !HAVE_STRCHR && !defined _LIBC
73 # ifndef strchr
74 # define strchr index
75 # endif
76 #endif
78 #if defined HAVE_UNISTD_H || defined _LIBC
79 # include <unistd.h>
80 #endif
82 #if defined HAVE_LOCALE_H || defined _LIBC
83 # include <locale.h>
84 #endif
86 #if defined HAVE_SYS_PARAM_H || defined _LIBC
87 # include <sys/param.h>
88 #endif
90 #include "gettext.h"
91 #include "gettextP.h"
92 #ifdef _LIBC
93 # include <libintl.h>
94 #else
95 # include "libgettext.h"
96 #endif
97 #include "hash-string.h"
99 /* Thread safetyness. */
100 #ifdef _LIBC
101 # include <bits/libc-lock.h>
102 #endif
104 /* @@ end of prolog @@ */
106 #ifdef _LIBC
107 /* Rename the non ANSI C functions. This is required by the standard
108 because some ANSI C functions will require linking with this object
109 file and the name space must not be polluted. */
110 # define getcwd __getcwd
111 # ifndef stpcpy
112 # define stpcpy __stpcpy
113 # endif
114 #else
115 # if !defined HAVE_GETCWD
116 char *getwd ();
117 # define getcwd(buf, max) getwd (buf)
118 # else
119 char *getcwd ();
120 # endif
121 # ifndef HAVE_STPCPY
122 static char *stpcpy PARAMS ((char *dest, const char *src));
123 # endif
124 #endif
126 /* Amount to increase buffer size by in each try. */
127 #define PATH_INCR 32
129 /* The following is from pathmax.h. */
130 /* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define
131 PATH_MAX but might cause redefinition warnings when sys/param.h is
132 later included (as on MORE/BSD 4.3). */
133 #if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && !defined(__GNUC__))
134 # include <limits.h>
135 #endif
137 #ifndef _POSIX_PATH_MAX
138 # define _POSIX_PATH_MAX 255
139 #endif
141 #if !defined(PATH_MAX) && defined(_PC_PATH_MAX)
142 # define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX))
143 #endif
145 /* Don't include sys/param.h if it already has been. */
146 #if defined(HAVE_SYS_PARAM_H) && !defined(PATH_MAX) && !defined(MAXPATHLEN)
147 # include <sys/param.h>
148 #endif
150 #if !defined(PATH_MAX) && defined(MAXPATHLEN)
151 # define PATH_MAX MAXPATHLEN
152 #endif
154 #ifndef PATH_MAX
155 # define PATH_MAX _POSIX_PATH_MAX
156 #endif
158 /* XPG3 defines the result of `setlocale (category, NULL)' as:
159 ``Directs `setlocale()' to query `category' and return the current
160 setting of `local'.''
161 However it does not specify the exact format. And even worse: POSIX
162 defines this not at all. So we can use this feature only on selected
163 system (e.g. those using GNU C Library). */
164 #ifdef _LIBC
165 # define HAVE_LOCALE_NULL
166 #endif
168 /* Name of the default domain used for gettext(3) prior any call to
169 textdomain(3). The default value for this is "messages". */
170 const char _nl_default_default_domain[] = "messages";
172 /* Value used as the default domain for gettext(3). */
173 const char *_nl_current_default_domain = _nl_default_default_domain;
175 /* Contains the default location of the message catalogs. */
176 const char _nl_default_dirname[] = GNULOCALEDIR;
178 /* List with bindings of specific domains created by bindtextdomain()
179 calls. */
180 struct binding *_nl_domain_bindings;
182 /* Prototypes for local functions. */
183 static const char *category_to_name PARAMS ((int category)) internal_function;
184 static const char *guess_category_value PARAMS ((int category,
185 const char *categoryname))
186 internal_function;
189 /* For those loosing systems which don't have `alloca' we have to add
190 some additional code emulating it. */
191 #ifdef HAVE_ALLOCA
192 /* Nothing has to be done. */
193 # define ADD_BLOCK(list, address) /* nothing */
194 # define FREE_BLOCKS(list) /* nothing */
195 #else
196 struct block_list
198 void *address;
199 struct block_list *next;
201 # define ADD_BLOCK(list, addr) \
202 do { \
203 struct block_list *newp = (struct block_list *) malloc (sizeof (*newp)); \
204 /* If we cannot get a free block we cannot add the new element to \
205 the list. */ \
206 if (newp != NULL) { \
207 newp->address = (addr); \
208 newp->next = (list); \
209 (list) = newp; \
211 } while (0)
212 # define FREE_BLOCKS(list) \
213 do { \
214 while (list != NULL) { \
215 struct block_list *old = list; \
216 list = list->next; \
217 free (old); \
219 } while (0)
220 # undef alloca
221 # define alloca(size) (malloc (size))
222 #endif /* have alloca */
225 /* Names for the libintl functions are a problem. They must not clash
226 with existing names and they should follow ANSI C. But this source
227 code is also used in GNU C Library where the names have a __
228 prefix. So we have to make a difference here. */
229 #ifdef _LIBC
230 # define DCGETTEXT __dcgettext
231 #else
232 # define DCGETTEXT dcgettext__
233 #endif
235 /* Checking whether the binaries runs SUID must be done and glibc provides
236 easier methods therefore we make a difference here. */
237 #ifdef _LIBC
238 # define ENABLE_SECURE __libc_enable_secure
239 # define DETERMINE_SECURE
240 #else
241 static int enable_secure;
242 # define ENABLE_SECURE (enable_secure == 1)
243 # define DETERMINE_SECURE \
244 if (enable_secure == 0) \
246 if (getuid () != geteuid () || getgid () != getegid ()) \
247 enable_secure = 1; \
248 else \
249 enable_secure = -1; \
251 #endif
253 /* Look up MSGID in the DOMAINNAME message catalog for the current CATEGORY
254 locale. */
255 char *
256 DCGETTEXT (domainname, msgid, category)
257 const char *domainname;
258 const char *msgid;
259 int category;
261 #ifndef HAVE_ALLOCA
262 struct block_list *block_list = NULL;
263 #endif
264 struct loaded_l10nfile *domain;
265 struct binding *binding;
266 const char *categoryname;
267 const char *categoryvalue;
268 char *dirname, *xdomainname;
269 char *single_locale;
270 char *retval;
271 int saved_errno = errno;
273 /* If no real MSGID is given return NULL. */
274 if (msgid == NULL)
275 return NULL;
277 /* See whether this is a SUID binary or not. */
278 DETERMINE_SECURE;
280 /* If DOMAINNAME is NULL, we are interested in the default domain. If
281 CATEGORY is not LC_MESSAGES this might not make much sense but the
282 definition left this undefined. */
283 if (domainname == NULL)
284 domainname = _nl_current_default_domain;
286 /* First find matching binding. */
287 for (binding = _nl_domain_bindings; binding != NULL; binding = binding->next)
289 int compare = strcmp (domainname, binding->domainname);
290 if (compare == 0)
291 /* We found it! */
292 break;
293 if (compare < 0)
295 /* It is not in the list. */
296 binding = NULL;
297 break;
301 if (binding == NULL)
302 dirname = (char *) _nl_default_dirname;
303 else if (binding->dirname[0] == '/')
304 dirname = binding->dirname;
305 else
307 /* We have a relative path. Make it absolute now. */
308 size_t dirname_len = strlen (binding->dirname) + 1;
309 size_t path_max;
310 char *ret;
312 path_max = (unsigned int) PATH_MAX;
313 path_max += 2; /* The getcwd docs say to do this. */
315 dirname = (char *) alloca (path_max + dirname_len);
316 ADD_BLOCK (block_list, dirname);
318 __set_errno (0);
319 while ((ret = getcwd (dirname, path_max)) == NULL && errno == ERANGE)
321 path_max += PATH_INCR;
322 dirname = (char *) alloca (path_max + dirname_len);
323 ADD_BLOCK (block_list, dirname);
324 __set_errno (0);
327 if (ret == NULL)
329 /* We cannot get the current working directory. Don't signal an
330 error but simply return the default string. */
331 FREE_BLOCKS (block_list);
332 __set_errno (saved_errno);
333 return (char *) msgid;
336 stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
339 /* Now determine the symbolic name of CATEGORY and its value. */
340 categoryname = category_to_name (category);
341 categoryvalue = guess_category_value (category, categoryname);
343 xdomainname = (char *) alloca (strlen (categoryname)
344 + strlen (domainname) + 5);
345 ADD_BLOCK (block_list, xdomainname);
347 stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
348 domainname),
349 ".mo");
351 /* Creating working area. */
352 single_locale = (char *) alloca (strlen (categoryvalue) + 1);
353 ADD_BLOCK (block_list, single_locale);
356 /* Search for the given string. This is a loop because we perhaps
357 got an ordered list of languages to consider for the translation. */
358 while (1)
360 /* Make CATEGORYVALUE point to the next element of the list. */
361 while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
362 ++categoryvalue;
363 if (categoryvalue[0] == '\0')
365 /* The whole contents of CATEGORYVALUE has been searched but
366 no valid entry has been found. We solve this situation
367 by implicitly appending a "C" entry, i.e. no translation
368 will take place. */
369 single_locale[0] = 'C';
370 single_locale[1] = '\0';
372 else
374 char *cp = single_locale;
375 while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
376 *cp++ = *categoryvalue++;
377 *cp = '\0';
379 /* When this is a SUID binary we must not allow accessing files
380 outside the dedicated directories. */
381 if (ENABLE_SECURE
382 && (memchr (single_locale, '/',
383 _nl_find_language (single_locale) - single_locale)
384 != NULL))
385 /* Ingore this entry. */
386 continue;
389 /* If the current locale value is C (or POSIX) we don't load a
390 domain. Return the MSGID. */
391 if (strcmp (single_locale, "C") == 0
392 || strcmp (single_locale, "POSIX") == 0)
394 FREE_BLOCKS (block_list);
395 __set_errno (saved_errno);
396 return (char *) msgid;
400 /* Find structure describing the message catalog matching the
401 DOMAINNAME and CATEGORY. */
402 domain = _nl_find_domain (dirname, single_locale, xdomainname);
404 if (domain != NULL)
406 retval = _nl_find_msg (domain, msgid);
408 if (retval == NULL)
410 int cnt;
412 for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
414 retval = _nl_find_msg (domain->successor[cnt], msgid);
416 if (retval != NULL)
417 break;
421 if (retval != NULL)
423 FREE_BLOCKS (block_list);
424 __set_errno (saved_errno);
425 return retval;
429 /* NOTREACHED */
432 #ifdef _LIBC
433 /* Alias for function name in GNU C Library. */
434 weak_alias (__dcgettext, dcgettext);
435 #endif
438 char *
439 internal_function
440 _nl_find_msg (domain_file, msgid)
441 struct loaded_l10nfile *domain_file;
442 const char *msgid;
444 size_t act = 0;
445 size_t top, bottom;
446 struct loaded_domain *domain;
448 if (domain_file->decided == 0)
449 _nl_load_domain (domain_file);
451 if (domain_file->data == NULL)
452 return NULL;
454 domain = (struct loaded_domain *) domain_file->data;
456 /* Locate the MSGID and its translation. */
457 if (domain->hash_size > 2 && domain->hash_tab != NULL)
459 /* Use the hashing table. */
460 nls_uint32 len = strlen (msgid);
461 nls_uint32 hash_val = hash_string (msgid);
462 nls_uint32 idx = hash_val % domain->hash_size;
463 nls_uint32 incr = 1 + (hash_val % (domain->hash_size - 2));
464 nls_uint32 nstr = W (domain->must_swap, domain->hash_tab[idx]);
466 if (nstr == 0)
467 /* Hash table entry is empty. */
468 return NULL;
470 if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
471 && strcmp (msgid,
472 domain->data + W (domain->must_swap,
473 domain->orig_tab[nstr - 1].offset)) == 0)
475 /* We found an entry. If we have to convert the string to use
476 a different character set this is the time. */
477 char *result =
478 (char *) domain->data + W (domain->must_swap,
479 domain->trans_tab[nstr - 1].offset);
481 if (
482 #ifdef _LIBC
483 domain->conv != (__gconv_t) -1
484 #else
485 # if HAVE_ICONV
486 domain->conv != (iconv_t) -1
487 # endif
488 #endif
491 /* We are supposed to do a conversion. First allocate an
492 appropriate table with the same structure as the hash
493 table in the file where we can put the pointers to the
494 converted strings in. */
495 if (domain->conv_tab == NULL
496 && ((domain->conv_tab = (char **) calloc (domain->hash_size,
497 sizeof (char *)))
498 == NULL))
499 /* Mark that we didn't succeed allocating a table. */
500 domain->conv_tab = (char **) -1;
502 if (domain->conv_tab == (char **) -1)
503 /* Nothing we can do, no more memory. */
504 return NULL;
506 if (domain->conv_tab[idx] == NULL)
508 /* We haven't used this string so far, so it is not
509 translated yet. Do this now. */
510 #ifdef _LIBC
511 /* For glibc we use a bit more efficient memory handling.
512 We allocate always larger blocks which get used over
513 time. This is faster than many small allocations. */
514 __libc_lock_define_initialized (static, lock)
515 static unsigned char *freemem;
516 static size_t freemem_size;
517 /* Note that we include the NUL byte. */
518 size_t resultlen = strlen (result) + 1;
519 const unsigned char *inbuf = result;
520 unsigned char *outbuf = freemem;
521 size_t written;
522 int res;
524 __libc_lock_lock (lock);
526 while ((res = __gconv (domain->conv,
527 &inbuf, inbuf + resultlen,
528 &outbuf, outbuf + freemem_size,
529 &written)) == __GCONV_OK)
531 if (res != __GCONV_FULL_OUTPUT)
532 goto out;
534 /* We must resize the buffer. */
535 freemem_size = MAX (2 * freemem_size, 4064);
536 freemem = (char *) malloc (freemem_size);
537 if (freemem == NULL)
538 goto out;
540 inbuf = result;
541 outbuf = freemem;
544 /* We have now in our buffer a converted string. Put this
545 in the hash table */
546 domain->conv_tab[idx] = freemem;
547 freemem_size -= outbuf - freemem;
548 freemem = outbuf;
550 out:
551 __libc_lock_unlock (lock);
552 #endif
555 result = domain->conv_tab[idx];
558 return result;
561 while (1)
563 if (idx >= domain->hash_size - incr)
564 idx -= domain->hash_size - incr;
565 else
566 idx += incr;
568 nstr = W (domain->must_swap, domain->hash_tab[idx]);
569 if (nstr == 0)
570 /* Hash table entry is empty. */
571 return NULL;
573 if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
574 && strcmp (msgid,
575 domain->data + W (domain->must_swap,
576 domain->orig_tab[nstr - 1].offset))
577 == 0)
578 return (char *) domain->data
579 + W (domain->must_swap, domain->trans_tab[nstr - 1].offset);
581 /* NOTREACHED */
584 /* Now we try the default method: binary search in the sorted
585 array of messages. */
586 bottom = 0;
587 top = domain->nstrings;
588 while (bottom < top)
590 int cmp_val;
592 act = (bottom + top) / 2;
593 cmp_val = strcmp (msgid, domain->data
594 + W (domain->must_swap,
595 domain->orig_tab[act].offset));
596 if (cmp_val < 0)
597 top = act;
598 else if (cmp_val > 0)
599 bottom = act + 1;
600 else
601 break;
604 /* If an translation is found return this. */
605 return bottom >= top ? NULL : (char *) domain->data
606 + W (domain->must_swap,
607 domain->trans_tab[act].offset);
611 /* Return string representation of locale CATEGORY. */
612 static const char *
613 internal_function
614 category_to_name (category)
615 int category;
617 const char *retval;
619 switch (category)
621 #ifdef LC_COLLATE
622 case LC_COLLATE:
623 retval = "LC_COLLATE";
624 break;
625 #endif
626 #ifdef LC_CTYPE
627 case LC_CTYPE:
628 retval = "LC_CTYPE";
629 break;
630 #endif
631 #ifdef LC_MONETARY
632 case LC_MONETARY:
633 retval = "LC_MONETARY";
634 break;
635 #endif
636 #ifdef LC_NUMERIC
637 case LC_NUMERIC:
638 retval = "LC_NUMERIC";
639 break;
640 #endif
641 #ifdef LC_TIME
642 case LC_TIME:
643 retval = "LC_TIME";
644 break;
645 #endif
646 #ifdef LC_MESSAGES
647 case LC_MESSAGES:
648 retval = "LC_MESSAGES";
649 break;
650 #endif
651 #ifdef LC_RESPONSE
652 case LC_RESPONSE:
653 retval = "LC_RESPONSE";
654 break;
655 #endif
656 #ifdef LC_ALL
657 case LC_ALL:
658 /* This might not make sense but is perhaps better than any other
659 value. */
660 retval = "LC_ALL";
661 break;
662 #endif
663 default:
664 /* If you have a better idea for a default value let me know. */
665 retval = "LC_XXX";
668 return retval;
671 /* Guess value of current locale from value of the environment variables. */
672 static const char *
673 internal_function
674 guess_category_value (category, categoryname)
675 int category;
676 const char *categoryname;
678 const char *retval;
680 /* The highest priority value is the `LANGUAGE' environment
681 variable. This is a GNU extension. */
682 retval = getenv ("LANGUAGE");
683 if (retval != NULL && retval[0] != '\0')
684 return retval;
686 /* `LANGUAGE' is not set. So we have to proceed with the POSIX
687 methods of looking to `LC_ALL', `LC_xxx', and `LANG'. On some
688 systems this can be done by the `setlocale' function itself. */
689 #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
690 return setlocale (category, NULL);
691 #else
692 /* Setting of LC_ALL overwrites all other. */
693 retval = getenv ("LC_ALL");
694 if (retval != NULL && retval[0] != '\0')
695 return retval;
697 /* Next comes the name of the desired category. */
698 retval = getenv (categoryname);
699 if (retval != NULL && retval[0] != '\0')
700 return retval;
702 /* Last possibility is the LANG environment variable. */
703 retval = getenv ("LANG");
704 if (retval != NULL && retval[0] != '\0')
705 return retval;
707 /* We use C as the default domain. POSIX says this is implementation
708 defined. */
709 return "C";
710 #endif
713 /* @@ begin of epilog @@ */
715 /* We don't want libintl.a to depend on any other library. So we
716 avoid the non-standard function stpcpy. In GNU C Library this
717 function is available, though. Also allow the symbol HAVE_STPCPY
718 to be defined. */
719 #if !_LIBC && !HAVE_STPCPY
720 static char *
721 stpcpy (dest, src)
722 char *dest;
723 const char *src;
725 while ((*dest++ = *src++) != '\0')
726 /* Do nothing. */ ;
727 return dest - 1;
729 #endif
732 #ifdef _LIBC
733 /* If we want to free all resources we have to do some work at
734 program's end. */
735 static void __attribute__ ((unused))
736 free_mem (void)
738 struct binding *runp;
740 for (runp = _nl_domain_bindings; runp != NULL; runp = runp->next)
742 free (runp->domainname);
743 if (runp->dirname != _nl_default_dirname)
744 /* Yes, this is a pointer comparison. */
745 free (runp->dirname);
748 if (_nl_current_default_domain != _nl_default_default_domain)
749 /* Yes, again a pointer comparison. */
750 free ((char *) _nl_current_default_domain);
753 text_set_element (__libc_subfreeres, free_mem);
754 #endif