intl/dcgettext.c

   1 /* dcgettext.c -- implementation of the dcgettext(3) function
   2    Copyright (C) 1995, 1996 Free Software Foundation, Inc.
   3
   4 This file is part of the GNU C Library.  Its master source is NOT part of
   5 the C library, however.  The master source lives in /gd/gnu/lib.
   6
   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.
  11
  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.
  16
  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
  19 not, write to the Free Software Foundation, Inc., 675 Mass Ave,
  20 Cambridge, MA 02139, USA.  */
  21
  22 #ifdef HAVE_CONFIG_H
  23 # include <config.h>
  24 #endif
  25
  26 #include <sys/types.h>
  27
  28 #ifdef __GNUC__
  29 # define alloca __builtin_alloca
  30 #else
  31 # if defined HAVE_ALLOCA_H || defined _LIBC
  32 #  include <alloca.h>
  33 # else
  34 #  ifdef _AIX
  35  #pragma alloca
  36 #  else
  37 #   ifndef alloca
  38 char *alloca ();
  39 #   endif
  40 #  endif
  41 # endif
  42 #endif
  43
  44 #include <errno.h>
  45 #ifndef errno
  46 extern int errno;
  47 #endif
  48
  49 #if defined STDC_HEADERS || defined _LIBC
  50 # include <stdlib.h>
  51 #else
  52 char *getenv ();
  53 # ifdef HAVE_MALLOC_H
  54 #  include <malloc.h>
  55 # else
  56 void free ();
  57 # endif
  58 #endif
  59
  60 #if defined HAVE_STRING_H || defined _LIBC
  61 # include <string.h>
  62 #else
  63 # include <strings.h>
  64 #endif
  65 #if !HAVE_STRCHR && !defined _LIBC
  66 # ifndef strchr
  67 #  define strchr index
  68 # endif
  69 #endif
  70
  71 #if defined HAVE_UNISTD_H || defined _LIBC
  72 # include <unistd.h>
  73 #endif
  74
  75 #include "gettext.h"
  76 #include "gettextP.h"
  77 #ifdef _LIBC
  78 # include <libintl.h>
  79 #else
  80 # include "libgettext.h"
  81 #endif
  82 #include "hash-string.h"
  83
  84 /* @@ end of prolog @@ */
  85
  86 #ifdef _LIBC
  87 /* Rename the non ANSI C functions.  This is required by the standard
  88    because some ANSI C functions will require linking with this object
  89    file and the name space must not be polluted.  */
  90 # define getcwd __getcwd
  91 # define stpcpy __stpcpy
  92 #else
  93 # if !defined HAVE_GETCWD
  94 char *getwd ();
  95 #  define getcwd(buf, max) getwd (buf)
  96 # else
  97 char *getcwd ();
  98 # endif
  99 #endif
 100
 101 /* Amount to increase buffer size by in each try.  */
 102 #define PATH_INCR 32
 103
 104 /* The following is from pathmax.h.  */
 105 /* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define
 106    PATH_MAX but might cause redefinition warnings when sys/param.h is
 107    later included (as on MORE/BSD 4.3).  */
 108 #if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && !defined(__GNUC__))
 109 # include <limits.h>
 110 #endif
 111
 112 #ifndef _POSIX_PATH_MAX
 113 # define _POSIX_PATH_MAX 255
 114 #endif
 115
 116 #if !defined(PATH_MAX) && defined(_PC_PATH_MAX)
 117 # define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX))
 118 #endif
 119
 120 /* Don't include sys/param.h if it already has been.  */
 121 #if defined(HAVE_SYS_PARAM_H) && !defined(PATH_MAX) && !defined(MAXPATHLEN)
 122 # include <sys/param.h>
 123 #endif
 124
 125 #if !defined(PATH_MAX) && defined(MAXPATHLEN)
 126 # define PATH_MAX MAXPATHLEN
 127 #endif
 128
 129 #ifndef PATH_MAX
 130 # define PATH_MAX _POSIX_PATH_MAX
 131 #endif
 132
 133 /* XPG3 defines the result of `setlocale (category, NULL)' as:
 134    ``Directs `setlocale()' to query `category' and return the current
 135      setting of `local'.''
 136    However it does not specify the exact format.  And even worse: POSIX
 137    defines this not at all.  So we can use this feature only on selected
 138    system (e.g. those using GNU C Library).  */
 139 #ifdef _LIBC
 140 # define HAVE_LOCALE_NULL
 141 #endif
 142
 143 /* Name of the default domain used for gettext(3) prior any call to
 144    textdomain(3).  The default value for this is "messages".  */
 145 const char _nl_default_default_domain[] = "messages";
 146
 147 /* Value used as the default domain for gettext(3).  */
 148 const char *_nl_current_default_domain = _nl_default_default_domain;
 149
 150 /* Contains the default location of the message catalogs.  */
 151 const char _nl_default_dirname[] = GNULOCALEDIR;
 152
 153 /* List with bindings of specific domains created by bindtextdomain()
 154    calls.  */
 155 struct binding *_nl_domain_bindings;
 156
 157 /* Prototypes for local functions.  */
 158 static char *find_msg PARAMS ((struct loaded_l10nfile *domain_file,
 159                                const char *msgid));
 160 static const char *category_to_name PARAMS ((int category));
 161 static const char *guess_category_value PARAMS ((int category,
 162                                                  const char *categoryname));
 163
 164
 165 #ifdef _LIBC
 166 #define HAVE_ALLOCA 1
 167 #endif
 168
 169 /* For those loosing systems which don't have `alloca' we have to add
 170    some additional code emulating it.  */
 171 #ifdef HAVE_ALLOCA
 172 /* Nothing has to be done.  */
 173 # define ADD_BLOCK(list, address) /* nothing */
 174 # define FREE_BLOCKS(list) /* nothing */
 175 #else
 176 struct block_list
 177 {
 178   void *address;
 179   struct block_list *next;
 180 };
 181 # define ADD_BLOCK(list, addr)                                                \
 182   do {                                                                        \
 183     struct block_list *newp = (struct block_list *) malloc (sizeof (*newp));  \
 184     /* If we cannot get a free block we cannot add the new element to         \
 185        the list.  */                                                          \
 186     if (newp != NULL) {                                                       \
 187       newp->address = (addr);                                                 \
 188       newp->next = (list);                                                    \
 189       (list) = newp;                                                          \
 190     }                                                                         \
 191   } while (0)
 192 # define FREE_BLOCKS(list)                                                    \
 193   do {                                                                        \
 194     while (list != NULL) {                                                    \
 195       struct block_list *old = list;                                          \
 196       list = list->next;                                                      \
 197       free (old);                                                             \
 198     }                                                                         \
 199   } while (0)
 200 # undef alloca
 201 # define alloca(size) (malloc (size))
 202 #endif  /* have alloca */
 203
 204
 205 /* Names for the libintl functions are a problem.  They must not clash
 206    with existing names and they should follow ANSI C.  But this source
 207    code is also used in GNU C Library where the names have a __
 208    prefix.  So we have to make a difference here.  */
 209 #ifdef _LIBC
 210 # define DCGETTEXT __dcgettext
 211 #else
 212 # define DCGETTEXT dcgettext__
 213 #endif
 214
 215 /* Look up MSGID in the DOMAINNAME message catalog for the current CATEGORY
 216    locale.  */
 217 char *
 218 DCGETTEXT (domainname, msgid, category)
 219      const char *domainname;
 220      const char *msgid;
 221      int category;
 222 {
 223 #ifndef HAVE_ALLOCA
 224   struct block_list *alloca_list = NULL;
 225 #endif
 226   struct loaded_l10nfile *domain;
 227   struct binding *binding;
 228   const char *categoryname;
 229   const char *categoryvalue;
 230   char *dirname, *xdomainname;
 231   char *single_locale;
 232   char *retval;
 233   int saved_errno = errno;
 234
 235   /* If no real MSGID is given return NULL.  */
 236   if (msgid == NULL)
 237     return NULL;
 238
 239   /* If DOMAINNAME is NULL, we are interested in the default domain.  If
 240      CATEGORY is not LC_MESSAGES this might not make much sense but the
 241      defintion left this undefined.  */
 242   if (domainname == NULL)
 243     domainname = _nl_current_default_domain;
 244
 245   /* First find matching binding.  */
 246   for (binding = _nl_domain_bindings; binding != NULL; binding = binding->next)
 247     {
 248       int compare = strcmp (domainname, binding->domainname);
 249       if (compare == 0)
 250         /* We found it!  */
 251         break;
 252       if (compare < 0)
 253         {
 254           /* It is not in the list.  */
 255           binding = NULL;
 256           break;
 257         }
 258     }
 259
 260   if (binding == NULL)
 261     dirname = (char *) _nl_default_dirname;
 262   else if (binding->dirname[0] == '/')
 263     dirname = binding->dirname;
 264   else
 265     {
 266       /* We have a relative path.  Make it absolute now.  */
 267       size_t dirname_len = strlen (binding->dirname) + 1;
 268       size_t path_max;
 269       char *ret;
 270
 271       path_max = (unsigned) PATH_MAX;
 272       path_max += 2;            /* The getcwd docs say to do this.  */
 273
 274       dirname = (char *) alloca (path_max + dirname_len);
 275       ADD_BLOCK (block_list, dirname);
 276
 277       errno = 0;
 278       while ((ret = getcwd (dirname, path_max)) == NULL && errno == ERANGE)
 279         {
 280           path_max += PATH_INCR;
 281           dirname = (char *) alloca (path_max + dirname_len);
 282           ADD_BLOCK (block_list, dirname);
 283           errno = 0;
 284         }
 285
 286       if (ret == NULL)
 287         {
 288           /* We cannot get the current working directory.  Don't signal an
 289              error but simply return the default string.  */
 290           FREE_BLOCKS (block_list);
 291           errno = saved_errno;
 292           return (char *) msgid;
 293         }
 294
 295       /* We don't want libintl.a to depend on any other library.  So
 296          we avoid the non-standard function stpcpy.  In GNU C Library
 297          this function is available, though.  Also allow the symbol
 298          HAVE_STPCPY to be defined.  */
 299 #if defined _LIBC || defined HAVE_STPCPY
 300       stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
 301 #else
 302       strcat (dirname, "/");
 303       strcat (dirname, binding->dirname);
 304 #endif
 305     }
 306
 307   /* Now determine the symbolic name of CATEGORY and its value.  */
 308   categoryname = category_to_name (category);
 309   categoryvalue = guess_category_value (category, categoryname);
 310
 311   xdomainname = (char *) alloca (strlen (categoryname)
 312                                  + strlen (domainname) + 5);
 313   ADD_BLOCK (block_list, xdomainname);
 314   /* We don't want libintl.a to depend on any other library.  So we
 315      avoid the non-standard function stpcpy.  In GNU C Library this
 316      function is available, though.  Also allow the symbol HAVE_STPCPY
 317      to be defined.  */
 318 #if defined _LIBC || defined HAVE_STPCPY
 319   stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
 320                   domainname),
 321           ".mo");
 322 #else
 323   strcpy (xdomainname, categoryname);
 324   strcat (xdomainname, "/");
 325   strcat (xdomainname, domainname);
 326   strcat (xdomainname, ".mo");
 327 #endif
 328
 329   /* Creating working area.  */
 330   single_locale = (char *) alloca (strlen (categoryvalue) + 1);
 331   ADD_BLOCK (block_list, single_locale);
 332
 333
 334   /* Search for the given string.  This is a loop because we perhaps
 335      got an ordered list of languages to consider for th translation.  */
 336   while (1)
 337     {
 338       /* Make CATEGORYVALUE point to the next element of the list.  */
 339       while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
 340         ++categoryvalue;
 341       if (categoryvalue[0] == '\0')
 342         {
 343           /* The whole contents of CATEGORYVALUE has been searched but
 344              no valid entry has been found.  We solve this situation
 345              by implicitely appending a "C" entry, i.e. no translation
 346              will take place.  */
 347           single_locale[0] = 'C';
 348           single_locale[1] = '\0';
 349         }
 350       else
 351         {
 352           char *cp = single_locale;
 353           while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
 354             *cp++ = *categoryvalue++;
 355           *cp = '\0';
 356         }
 357
 358       /* If the current locale value is C (or POSIX) we don't load a
 359          domain.  Return the MSGID.  */
 360       if (strcmp (single_locale, "C") == 0
 361           || strcmp (single_locale, "POSIX") == 0)
 362         {
 363           FREE_BLOCKS (block_list);
 364           errno = saved_errno;
 365           return (char *) msgid;
 366         }
 367
 368
 369       /* Find structure describing the message catalog matching the
 370          DOMAINNAME and CATEGORY.  */
 371       domain = _nl_find_domain (dirname, single_locale, xdomainname);
 372
 373       if (domain != NULL)
 374         {
 375           retval = find_msg (domain, msgid);
 376
 377           if (retval == NULL)
 378             {
 379               int cnt;
 380
 381               for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
 382                 {
 383                   retval = find_msg (domain->successor[cnt], msgid);
 384
 385                   if (retval != NULL)
 386                     break;
 387                 }
 388             }
 389
 390           if (retval != NULL)
 391             {
 392               FREE_BLOCKS (block_list);
 393               errno = saved_errno;
 394               return retval;
 395             }
 396         }
 397     }
 398   /* NOTREACHED */
 399 }
 400
 401 #ifdef _LIBC
 402 /* Alias for function name in GNU C Library.  */
 403 weak_alias (__dcgettext, dcgettext);
 404 #endif
 405
 406
 407 static char *
 408 find_msg (domain_file, msgid)
 409      struct loaded_l10nfile *domain_file;
 410      const char *msgid;
 411 {
 412   size_t top, act, bottom;
 413   struct loaded_domain *domain;
 414
 415   if (domain_file->decided == 0)
 416     _nl_load_domain (domain_file);
 417
 418   if (domain_file->data == NULL)
 419     return NULL;
 420
 421   domain = (struct loaded_domain *) domain_file->data;
 422
 423   /* Locate the MSGID and its translation.  */
 424   if (domain->hash_size > 2 && domain->hash_tab != NULL)
 425     {
 426       /* Use the hashing table.  */
 427       nls_uint32 len = strlen (msgid);
 428       nls_uint32 hash_val = hash_string (msgid);
 429       nls_uint32 idx = hash_val % domain->hash_size;
 430       nls_uint32 incr = 1 + (hash_val % (domain->hash_size - 2));
 431       nls_uint32 nstr = W (domain->must_swap, domain->hash_tab[idx]);
 432
 433       if (nstr == 0)
 434         /* Hash table entry is empty.  */
 435         return NULL;
 436
 437       if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 438           && strcmp (msgid,
 439                      domain->data + W (domain->must_swap,
 440                                        domain->orig_tab[nstr - 1].offset)) == 0)
 441         return (char *) domain->data + W (domain->must_swap,
 442                                           domain->trans_tab[nstr - 1].offset);
 443
 444       while (1)
 445         {
 446           if (idx >= domain->hash_size - incr)
 447             idx -= domain->hash_size - incr;
 448           else
 449             idx += incr;
 450
 451           nstr = W (domain->must_swap, domain->hash_tab[idx]);
 452           if (nstr == 0)
 453             /* Hash table entry is empty.  */
 454             return NULL;
 455
 456           if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 457               && strcmp (msgid,
 458                          domain->data + W (domain->must_swap,
 459                                            domain->orig_tab[nstr - 1].offset))
 460                  == 0)
 461             return (char *) domain->data
 462               + W (domain->must_swap, domain->trans_tab[nstr - 1].offset);
 463         }
 464       /* NOTREACHED */
 465     }
 466
 467   /* Now we try the default method:  binary search in the sorted
 468      array of messages.  */
 469   bottom = 0;
 470   top = domain->nstrings;
 471   while (bottom < top)
 472     {
 473       int cmp_val;
 474
 475       act = (bottom + top) / 2;
 476       cmp_val = strcmp (msgid, domain->data
 477                                + W (domain->must_swap,
 478                                     domain->orig_tab[act].offset));
 479       if (cmp_val < 0)
 480         top = act;
 481       else if (cmp_val > 0)
 482         bottom = act + 1;
 483       else
 484         break;
 485     }
 486
 487   /* If an translation is found return this.  */
 488   return bottom >= top ? NULL : (char *) domain->data
 489                                 + W (domain->must_swap,
 490                                      domain->trans_tab[act].offset);
 491 }
 492
 493
 494 /* Return string representation of locale CATEGORY.  */
 495 static const char *
 496 category_to_name (category)
 497      int category;
 498 {
 499   const char *retval;
 500
 501   switch (category)
 502   {
 503 #ifdef LC_COLLATE
 504   case LC_COLLATE:
 505     retval = "LC_COLLATE";
 506     break;
 507 #endif
 508 #ifdef LC_CTYPE
 509   case LC_CTYPE:
 510     retval = "LC_CTYPE";
 511     break;
 512 #endif
 513 #ifdef LC_MONETARY
 514   case LC_MONETARY:
 515     retval = "LC_MONETARY";
 516     break;
 517 #endif
 518 #ifdef LC_NUMERIC
 519   case LC_NUMERIC:
 520     retval = "LC_NUMERIC";
 521     break;
 522 #endif
 523 #ifdef LC_TIME
 524   case LC_TIME:
 525     retval = "LC_TIME";
 526     break;
 527 #endif
 528 #ifdef LC_MESSAGES
 529   case LC_MESSAGES:
 530     retval = "LC_MESSAGES";
 531     break;
 532 #endif
 533 #ifdef LC_RESPONSE
 534   case LC_RESPONSE:
 535     retval = "LC_RESPONSE";
 536     break;
 537 #endif
 538 #ifdef LC_ALL
 539   case LC_ALL:
 540     /* This might not make sense but is perhaps better than any other
 541        value.  */
 542     retval = "LC_ALL";
 543     break;
 544 #endif
 545   default:
 546     /* If you have a better idea for a default value let me know.  */
 547     retval = "LC_XXX";
 548   }
 549
 550   return retval;
 551 }
 552
 553 /* Guess value of current locale from value of the environment variables.  */
 554 static const char *guess_category_value (category, categoryname)
 555      int category;
 556      const char *categoryname;
 557 {
 558   const char *retval;
 559
 560   /* The highest priority value is the `LANGUAGE' environment
 561      variable.  This is a GNU extension.  */
 562   retval = getenv ("LANGUAGE");
 563   if (retval != NULL && retval[0] != '\0')
 564     return retval;
 565
 566   /* `LANGUAGE' is not set.  So we have to proceed with the POSIX
 567      methods of looking to `LC_ALL', `LC_xxx', and `LANG'.  On some
 568      systems this can be done by the `setlocale' function itself.  */
 569 #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
 570   return setlocale (category, NULL);
 571 #else
 572   /* Setting of LC_ALL overwrites all other.  */
 573   retval = getenv ("LC_ALL");
 574   if (retval != NULL && retval[0] != '\0')
 575     return retval;
 576
 577   /* Next comes the name of the desired category.  */
 578   retval = getenv (categoryname);
 579   if (retval != NULL && retval[0] != '\0')
 580     return retval;
 581
 582   /* Last possibility is the LANG environment variable.  */
 583   retval = getenv ("LANG");
 584   if (retval != NULL && retval[0] != '\0')
 585     return retval;
 586
 587   /* We use C as the default domain.  POSIX says this is implementation
 588      defined.  */
 589   return "C";
 590 #endif
 591 }