intl/dcgettext.c

   1 /* dcgettext.c -- implementation of the dcgettext(3) function
   2 Copyright (C) 1995, 1996 Free Software Foundation, Inc.
   3 Written by Ulrich Drepper <drepper@gnu.ai.mit.edu>, 1995.
   4
   5 The GNU C Library is free software; you can redistribute it and/or
   6 modify it under the terms of the GNU Library General Public License as
   7 published by the Free Software Foundation; either version 2 of the
   8 License, or (at your option) any later version.
   9
  10 The GNU C Library 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 the GNU C Library; see the file COPYING.LIB.  If
  17 not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  18 Boston, MA 02111-1307, USA.  */
  19
  20 #ifdef HAVE_CONFIG_H
  21 # include <config.h>
  22 #endif
  23
  24 #include <sys/types.h>
  25
  26 #ifdef __GNUC__
  27 # define alloca __builtin_alloca
  28 #else
  29 # if defined HAVE_ALLOCA_H || defined _LIBC
  30 #  include <alloca.h>
  31 # else
  32 #  ifdef _AIX
  33  #pragma alloca
  34 #  else
  35 #   ifndef alloca
  36 char *alloca ();
  37 #   endif
  38 #  endif
  39 # endif
  40 #endif
  41
  42 #include <errno.h>
  43 #ifndef errno
  44 extern int errno;
  45 #endif
  46
  47 #if defined STDC_HEADERS || defined _LIBC
  48 # include <stdlib.h>
  49 #else
  50 char *getenv ();
  51 # ifdef HAVE_MALLOC_H
  52 #  include <malloc.h>
  53 # else
  54 void free ();
  55 # endif
  56 #endif
  57
  58 #if defined HAVE_STRING_H || defined _LIBC
  59 # include <string.h>
  60 #else
  61 # include <strings.h>
  62 #endif
  63 #if !HAVE_STRCHR && !defined _LIBC
  64 # ifndef strchr
  65 #  define strchr index
  66 # endif
  67 #endif
  68
  69 #if defined HAVE_UNISTD_H || defined _LIBC
  70 # include <unistd.h>
  71 #endif
  72
  73 #include "gettext.h"
  74 #include "gettextP.h"
  75 #ifdef _LIBC
  76 # include <libintl.h>
  77 #else
  78 # include "libgettext.h"
  79 #endif
  80 #include "hash-string.h"
  81
  82 /* @@ end of prolog @@ */
  83
  84 #ifdef _LIBC
  85 /* Rename the non ANSI C functions.  This is required by the standard
  86    because some ANSI C functions will require linking with this object
  87    file and the name space must not be polluted.  */
  88 # define getcwd __getcwd
  89 # define stpcpy __stpcpy
  90 #endif
  91
  92 #if !defined HAVE_GETCWD && !defined _LIBC
  93 char *getwd ();
  94 # define getcwd(buf, max) getwd (buf)
  95 #else
  96 char *getcwd ();
  97 #endif
  98
  99 /* Amount to increase buffer size by in each try.  */
 100 #define PATH_INCR 32
 101
 102 /* The following is from pathmax.h.  */
 103 /* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define
 104    PATH_MAX but might cause redefinition warnings when sys/param.h is
 105    later included (as on MORE/BSD 4.3).  */
 106 #if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && !defined(__GNUC__))
 107 # include <limits.h>
 108 #endif
 109
 110 #ifndef _POSIX_PATH_MAX
 111 # define _POSIX_PATH_MAX 255
 112 #endif
 113
 114 #if !defined(PATH_MAX) && defined(_PC_PATH_MAX)
 115 # define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX))
 116 #endif
 117
 118 /* Don't include sys/param.h if it already has been.  */
 119 #if defined(HAVE_SYS_PARAM_H) && !defined(PATH_MAX) && !defined(MAXPATHLEN)
 120 # include <sys/param.h>
 121 #endif
 122
 123 #if !defined(PATH_MAX) && defined(MAXPATHLEN)
 124 # define PATH_MAX MAXPATHLEN
 125 #endif
 126
 127 #ifndef PATH_MAX
 128 # define PATH_MAX _POSIX_PATH_MAX
 129 #endif
 130
 131 /* XPG3 defines the result of `setlocale (category, NULL)' as:
 132    ``Directs `setlocale()' to query `category' and return the current
 133      setting of `local'.''
 134    However it does not specify the exact format.  And even worse: POSIX
 135    defines this not at all.  So we can use this feature only on selected
 136    system (e.g. those using GNU C Library).  */
 137 #ifdef _LIBC
 138 # define HAVE_LOCALE_NULL
 139 #endif
 140
 141 /* Name of the default domain used for gettext(3) prior any call to
 142    textdomain(3).  The default value for this is "messages".  */
 143 const char _nl_default_default_domain[] = "messages";
 144
 145 /* Value used as the default domain for gettext(3).  */
 146 const char *_nl_current_default_domain = _nl_default_default_domain;
 147
 148 /* Contains the default location of the message catalogs.  */
 149 const char _nl_default_dirname[] = GNULOCALEDIR;
 150
 151 /* List with bindings of specific domains created by bindtextdomain()
 152    calls.  */
 153 struct binding *_nl_domain_bindings;
 154
 155 /* Prototypes for local functions.  */
 156 static char *find_msg PARAMS ((struct loaded_l10nfile *domain_file,
 157                                const char *msgid));
 158 static const char *category_to_name PARAMS ((int category));
 159 static const char *guess_category_value PARAMS ((int category,
 160                                                  const char *categoryname));
 161
 162
 163 /* Names for the libintl functions are a problem.  They must not clash
 164    with existing names and they should follow ANSI C.  But this source
 165    code is also used in GNU C Library where the names have a __
 166    prefix.  So we have to make a difference here.  */
 167 #ifdef _LIBC
 168 # define DCGETTEXT __dcgettext
 169 #else
 170 # define DCGETTEXT dcgettext__
 171 #endif
 172
 173 /* Look up MSGID in the DOMAINNAME message catalog for the current CATEGORY
 174    locale.  */
 175 char *
 176 DCGETTEXT (domainname, msgid, category)
 177      const char *domainname;
 178      const char *msgid;
 179      int category;
 180 {
 181   struct loaded_l10nfile *domain;
 182   struct binding *binding;
 183   const char *categoryname;
 184   const char *categoryvalue;
 185   char *dirname, *xdomainname;
 186   char *single_locale;
 187   char *retval;
 188   int saved_errno = errno;
 189
 190   /* If no real MSGID is given return NULL.  */
 191   if (msgid == NULL)
 192     return NULL;
 193
 194   /* If DOMAINNAME is NULL, we are interested in the default domain.  If
 195      CATEGORY is not LC_MESSAGES this might not make much sense but the
 196      defintion left this undefined.  */
 197   if (domainname == NULL)
 198     domainname = _nl_current_default_domain;
 199
 200   /* First find matching binding.  */
 201   for (binding = _nl_domain_bindings; binding != NULL; binding = binding->next)
 202     {
 203       int compare = strcmp (domainname, binding->domainname);
 204       if (compare == 0)
 205         /* We found it!  */
 206         break;
 207       if (compare < 0)
 208         {
 209           /* It is not in the list.  */
 210           binding = NULL;
 211           break;
 212         }
 213     }
 214
 215   if (binding == NULL)
 216     dirname = (char *) _nl_default_dirname;
 217   else if (binding->dirname[0] == '/')
 218     dirname = binding->dirname;
 219   else
 220     {
 221       /* We have a relative path.  Make it absolute now.  */
 222       size_t dirname_len = strlen (binding->dirname) + 1;
 223       size_t path_max;
 224       char *ret;
 225
 226       path_max = (unsigned) PATH_MAX;
 227       path_max += 2;            /* The getcwd docs say to do this.  */
 228
 229       dirname = (char *) alloca (path_max + dirname_len);
 230
 231       errno = 0;
 232       while ((ret = getcwd (dirname, path_max)) == NULL && errno == ERANGE)
 233         {
 234           path_max += PATH_INCR;
 235           dirname = (char *) alloca (path_max + dirname_len);
 236           errno = 0;
 237         }
 238
 239       if (ret == NULL)
 240         {
 241           /* We cannot get the current working directory.  Don't signal an
 242              error but simply return the default string.  */
 243           errno = saved_errno;
 244           return (char *) msgid;
 245         }
 246
 247       /* We don't want libintl.a to depend on any other library.  So
 248          we avoid the non-standard function stpcpy.  In GNU C Library
 249          this function is available, though.  Also allow the symbol
 250          HAVE_STPCPY to be defined.  */
 251 #if defined _LIBC || defined HAVE_STPCPY
 252       stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
 253 #else
 254       strcat (dirname, "/");
 255       strcat (dirname, binding->dirname);
 256 #endif
 257     }
 258
 259   /* Now determine the symbolic name of CATEGORY and its value.  */
 260   categoryname = category_to_name (category);
 261   categoryvalue = guess_category_value (category, categoryname);
 262
 263   xdomainname = (char *) alloca (strlen (categoryname)
 264                                  + strlen (domainname) + 5);
 265   /* We don't want libintl.a to depend on any other library.  So we
 266      avoid the non-standard function stpcpy.  In GNU C Library this
 267      function is available, though.  Also allow the symbol HAVE_STPCPY
 268      to be defined.  */
 269 #if defined _LIBC || defined HAVE_STPCPY
 270   stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
 271                   domainname),
 272           ".mo");
 273 #else
 274   strcpy (xdomainname, categoryname);
 275   strcat (xdomainname, "/");
 276   strcat (xdomainname, domainname);
 277   strcat (xdomainname, ".mo");
 278 #endif
 279
 280   /* Creating working area.  */
 281   single_locale = (char *) alloca (strlen (categoryvalue) + 1);
 282
 283
 284   /* Search for the given string.  This is a loop because we perhaps
 285      got an ordered list of languages to consider for th translation.  */
 286   while (1)
 287     {
 288       /* Make CATEGORYVALUE point to the next element of the list.  */
 289       while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
 290         ++categoryvalue;
 291       if (categoryvalue[0] == '\0')
 292         {
 293           /* The whole contents of CATEGORYVALUE has been searched but
 294              no valid entry has been found.  We solve this situation
 295              by implicitely appending a "C" entry, i.e. no translation
 296              will take place.  */
 297           single_locale[0] = 'C';
 298           single_locale[1] = '\0';
 299         }
 300       else
 301         {
 302           char *cp = single_locale;
 303           while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
 304             *cp++ = *categoryvalue++;
 305           *cp = '\0';
 306         }
 307
 308       /* If the current locale value is C (or POSIX) we don't load a
 309          domain.  Return the MSGID.  */
 310       if (strcmp (single_locale, "C") == 0
 311           || strcmp (single_locale, "POSIX") == 0)
 312         {
 313           errno = saved_errno;
 314           return (char *) msgid;
 315         }
 316
 317
 318       /* Find structure describing the message catalog matching the
 319          DOMAINNAME and CATEGORY.  */
 320       domain = _nl_find_domain (dirname, single_locale, xdomainname);
 321
 322       if (domain != NULL)
 323         {
 324           retval = find_msg (domain, msgid);
 325
 326           if (retval == NULL)
 327             {
 328               int cnt;
 329
 330               for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
 331                 {
 332                   retval = find_msg (domain->successor[cnt], msgid);
 333
 334                   if (retval != NULL)
 335                     break;
 336                 }
 337             }
 338
 339           if (retval != NULL)
 340             {
 341               errno = saved_errno;
 342               return retval;
 343             }
 344         }
 345     }
 346   /* NOTREACHED */
 347 }
 348
 349 #ifdef _LIBC
 350 /* Alias for function name in GNU C Library.  */
 351 weak_alias (__dcgettext, dcgettext);
 352 #endif
 353
 354
 355 static char *
 356 find_msg (domain_file, msgid)
 357      struct loaded_l10nfile *domain_file;
 358      const char *msgid;
 359 {
 360   size_t top, act, bottom;
 361   struct loaded_domain *domain;
 362
 363   if (domain_file->decided == 0)
 364     _nl_load_domain (domain_file);
 365
 366   if (domain_file->data == NULL)
 367     return NULL;
 368
 369   domain = (struct loaded_domain *) domain_file->data;
 370
 371   /* Locate the MSGID and its translation.  */
 372   if (domain->hash_size > 2 && domain->hash_tab != NULL)
 373     {
 374       /* Use the hashing table.  */
 375       nls_uint32 len = strlen (msgid);
 376       nls_uint32 hash_val = hash_string (msgid);
 377       nls_uint32 idx = hash_val % domain->hash_size;
 378       nls_uint32 incr = 1 + (hash_val % (domain->hash_size - 2));
 379       nls_uint32 nstr = W (domain->must_swap, domain->hash_tab[idx]);
 380
 381       if (nstr == 0)
 382         /* Hash table entry is empty.  */
 383         return NULL;
 384
 385       if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 386           && strcmp (msgid,
 387                      domain->data + W (domain->must_swap,
 388                                        domain->orig_tab[nstr - 1].offset)) == 0)
 389         return (char *) domain->data + W (domain->must_swap,
 390                                           domain->trans_tab[nstr - 1].offset);
 391
 392       while (1)
 393         {
 394           if (idx >= domain->hash_size - incr)
 395             idx -= domain->hash_size - incr;
 396           else
 397             idx += incr;
 398
 399           nstr = W (domain->must_swap, domain->hash_tab[idx]);
 400           if (nstr == 0)
 401             /* Hash table entry is empty.  */
 402             return NULL;
 403
 404           if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 405               && strcmp (msgid,
 406                          domain->data + W (domain->must_swap,
 407                                            domain->orig_tab[nstr - 1].offset))
 408                  == 0)
 409             return (char *) domain->data
 410               + W (domain->must_swap, domain->trans_tab[nstr - 1].offset);
 411         }
 412       /* NOTREACHED */
 413     }
 414
 415   /* Now we try the default method:  binary search in the sorted
 416      array of messages.  */
 417   bottom = 0;
 418   top = domain->nstrings;
 419   while (bottom < top)
 420     {
 421       int cmp_val;
 422
 423       act = (bottom + top) / 2;
 424       cmp_val = strcmp (msgid, domain->data
 425                                + W (domain->must_swap,
 426                                     domain->orig_tab[act].offset));
 427       if (cmp_val < 0)
 428         top = act;
 429       else if (cmp_val > 0)
 430         bottom = act + 1;
 431       else
 432         break;
 433     }
 434
 435   /* If an translation is found return this.  */
 436   return bottom >= top ? NULL : (char *) domain->data
 437                                 + W (domain->must_swap,
 438                                      domain->trans_tab[act].offset);
 439 }
 440
 441
 442 /* Return string representation of locale CATEGORY.  */
 443 static const char *
 444 category_to_name (category)
 445      int category;
 446 {
 447   const char *retval;
 448
 449   switch (category)
 450   {
 451 #ifdef LC_COLLATE
 452   case LC_COLLATE:
 453     retval = "LC_COLLATE";
 454     break;
 455 #endif
 456 #ifdef LC_CTYPE
 457   case LC_CTYPE:
 458     retval = "LC_CTYPE";
 459     break;
 460 #endif
 461 #ifdef LC_MONETARY
 462   case LC_MONETARY:
 463     retval = "LC_MONETARY";
 464     break;
 465 #endif
 466 #ifdef LC_NUMERIC
 467   case LC_NUMERIC:
 468     retval = "LC_NUMERIC";
 469     break;
 470 #endif
 471 #ifdef LC_TIME
 472   case LC_TIME:
 473     retval = "LC_TIME";
 474     break;
 475 #endif
 476 #ifdef LC_MESSAGES
 477   case LC_MESSAGES:
 478     retval = "LC_MESSAGES";
 479     break;
 480 #endif
 481 #ifdef LC_RESPONSE
 482   case LC_RESPONSE:
 483     retval = "LC_RESPONSE";
 484     break;
 485 #endif
 486 #ifdef LC_ALL
 487   case LC_ALL:
 488     /* This might not make sense but is perhaps better than any other
 489        value.  */
 490     retval = "LC_ALL";
 491     break;
 492 #endif
 493   default:
 494     /* If you have a better idea for a default value let me know.  */
 495     retval = "LC_XXX";
 496   }
 497
 498   return retval;
 499 }
 500
 501 /* Guess value of current locale from value of the environment variables.  */
 502 static const char *guess_category_value (category, categoryname)
 503      int category;
 504      const char *categoryname;
 505 {
 506   const char *retval;
 507
 508   /* The highest priority value is the `LANGUAGE' environment
 509      variable.  This is a GNU extension.  */
 510   retval = getenv ("LANGUAGE");
 511   if (retval != NULL && retval[0] != '\0')
 512     return retval;
 513
 514   /* `LANGUAGE' is not set.  So we have to proceed with the POSIX
 515      methods of looking to `LC_ALL', `LC_xxx', and `LANG'.  On some
 516      systems this can be done by the `setlocale' function itself.  */
 517 #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
 518   return setlocale (category, NULL);
 519 #else
 520   /* Setting of LC_ALL overwrites all other.  */
 521   retval = getenv ("LC_ALL");
 522   if (retval != NULL && retval[0] != '\0')
 523     return retval;
 524
 525   /* Next comes the name of the desired category.  */
 526   retval = getenv (categoryname);
 527   if (retval != NULL && retval[0] != '\0')
 528     return retval;
 529
 530   /* Last possibility is the LANG environment variable.  */
 531   retval = getenv ("LANG");
 532   if (retval != NULL && retval[0] != '\0')
 533     return retval;
 534
 535   /* We use C as the default domain.  POSIX says this is implementation
 536      defined.  */
 537   return "C";
 538 #endif
 539 }