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 # define HAVE_ALLOCA 1
  31 #else
  32 # if defined HAVE_ALLOCA_H || defined _LIBC
  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
  44
  45 #include <errno.h>
  46 #ifndef errno
  47 extern int errno;
  48 #endif
  49
  50 #if defined STDC_HEADERS || defined _LIBC
  51 # include <stdlib.h>
  52 #else
  53 char *getenv ();
  54 # ifdef HAVE_MALLOC_H
  55 #  include <malloc.h>
  56 # else
  57 void free ();
  58 # endif
  59 #endif
  60
  61 #if defined HAVE_STRING_H || defined _LIBC
  62 # include <string.h>
  63 #else
  64 # include <strings.h>
  65 #endif
  66 #if !HAVE_STRCHR && !defined _LIBC
  67 # ifndef strchr
  68 #  define strchr index
  69 # endif
  70 #endif
  71
  72 #if defined HAVE_UNISTD_H || defined _LIBC
  73 # include <unistd.h>
  74 #endif
  75
  76 #include "gettext.h"
  77 #include "gettextP.h"
  78 #ifdef _LIBC
  79 # include <libintl.h>
  80 #else
  81 # include "libgettext.h"
  82 #endif
  83 #include "hash-string.h"
  84
  85 /* @@ end of prolog @@ */
  86
  87 #ifdef _LIBC
  88 /* Rename the non ANSI C functions.  This is required by the standard
  89    because some ANSI C functions will require linking with this object
  90    file and the name space must not be polluted.  */
  91 # define getcwd __getcwd
  92 # define stpcpy __stpcpy
  93 #else
  94 # if !defined HAVE_GETCWD
  95 char *getwd ();
  96 #  define getcwd(buf, max) getwd (buf)
  97 # else
  98 char *getcwd ();
  99 # endif
 100 # ifndef HAVE_STPCPY
 101 static char *stpcpy PARAMS ((char *dest, const char *src));
 102 # endif
 103 #endif
 104
 105 /* Amount to increase buffer size by in each try.  */
 106 #define PATH_INCR 32
 107
 108 /* The following is from pathmax.h.  */
 109 /* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define
 110    PATH_MAX but might cause redefinition warnings when sys/param.h is
 111    later included (as on MORE/BSD 4.3).  */
 112 #if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && !defined(__GNUC__))
 113 # include <limits.h>
 114 #endif
 115
 116 #ifndef _POSIX_PATH_MAX
 117 # define _POSIX_PATH_MAX 255
 118 #endif
 119
 120 #if !defined(PATH_MAX) && defined(_PC_PATH_MAX)
 121 # define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX))
 122 #endif
 123
 124 /* Don't include sys/param.h if it already has been.  */
 125 #if defined(HAVE_SYS_PARAM_H) && !defined(PATH_MAX) && !defined(MAXPATHLEN)
 126 # include <sys/param.h>
 127 #endif
 128
 129 #if !defined(PATH_MAX) && defined(MAXPATHLEN)
 130 # define PATH_MAX MAXPATHLEN
 131 #endif
 132
 133 #ifndef PATH_MAX
 134 # define PATH_MAX _POSIX_PATH_MAX
 135 #endif
 136
 137 /* XPG3 defines the result of `setlocale (category, NULL)' as:
 138    ``Directs `setlocale()' to query `category' and return the current
 139      setting of `local'.''
 140    However it does not specify the exact format.  And even worse: POSIX
 141    defines this not at all.  So we can use this feature only on selected
 142    system (e.g. those using GNU C Library).  */
 143 #ifdef _LIBC
 144 # define HAVE_LOCALE_NULL
 145 #endif
 146
 147 /* Name of the default domain used for gettext(3) prior any call to
 148    textdomain(3).  The default value for this is "messages".  */
 149 const char _nl_default_default_domain[] = "messages";
 150
 151 /* Value used as the default domain for gettext(3).  */
 152 const char *_nl_current_default_domain = _nl_default_default_domain;
 153
 154 /* Contains the default location of the message catalogs.  */
 155 const char _nl_default_dirname[] = GNULOCALEDIR;
 156
 157 /* List with bindings of specific domains created by bindtextdomain()
 158    calls.  */
 159 struct binding *_nl_domain_bindings;
 160
 161 /* Prototypes for local functions.  */
 162 static char *find_msg PARAMS ((struct loaded_l10nfile *domain_file,
 163                                const char *msgid));
 164 static const char *category_to_name PARAMS ((int category));
 165 static const char *guess_category_value PARAMS ((int category,
 166                                                  const char *categoryname));
 167
 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       stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
 300     }
 301
 302   /* Now determine the symbolic name of CATEGORY and its value.  */
 303   categoryname = category_to_name (category);
 304   categoryvalue = guess_category_value (category, categoryname);
 305
 306   xdomainname = (char *) alloca (strlen (categoryname)
 307                                  + strlen (domainname) + 5);
 308   ADD_BLOCK (block_list, xdomainname);
 309   /* We don't want libintl.a to depend on any other library.  So we
 310      avoid the non-standard function stpcpy.  In GNU C Library this
 311      function is available, though.  Also allow the symbol HAVE_STPCPY
 312      to be defined.  */
 313   stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
 314                   domainname),
 315           ".mo");
 316
 317   /* Creating working area.  */
 318   single_locale = (char *) alloca (strlen (categoryvalue) + 1);
 319   ADD_BLOCK (block_list, single_locale);
 320
 321
 322   /* Search for the given string.  This is a loop because we perhaps
 323      got an ordered list of languages to consider for th translation.  */
 324   while (1)
 325     {
 326       /* Make CATEGORYVALUE point to the next element of the list.  */
 327       while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
 328         ++categoryvalue;
 329       if (categoryvalue[0] == '\0')
 330         {
 331           /* The whole contents of CATEGORYVALUE has been searched but
 332              no valid entry has been found.  We solve this situation
 333              by implicitely appending a "C" entry, i.e. no translation
 334              will take place.  */
 335           single_locale[0] = 'C';
 336           single_locale[1] = '\0';
 337         }
 338       else
 339         {
 340           char *cp = single_locale;
 341           while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
 342             *cp++ = *categoryvalue++;
 343           *cp = '\0';
 344         }
 345
 346       /* If the current locale value is C (or POSIX) we don't load a
 347          domain.  Return the MSGID.  */
 348       if (strcmp (single_locale, "C") == 0
 349           || strcmp (single_locale, "POSIX") == 0)
 350         {
 351           FREE_BLOCKS (block_list);
 352           errno = saved_errno;
 353           return (char *) msgid;
 354         }
 355
 356
 357       /* Find structure describing the message catalog matching the
 358          DOMAINNAME and CATEGORY.  */
 359       domain = _nl_find_domain (dirname, single_locale, xdomainname);
 360
 361       if (domain != NULL)
 362         {
 363           retval = find_msg (domain, msgid);
 364
 365           if (retval == NULL)
 366             {
 367               int cnt;
 368
 369               for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
 370                 {
 371                   retval = find_msg (domain->successor[cnt], msgid);
 372
 373                   if (retval != NULL)
 374                     break;
 375                 }
 376             }
 377
 378           if (retval != NULL)
 379             {
 380               FREE_BLOCKS (block_list);
 381               errno = saved_errno;
 382               return retval;
 383             }
 384         }
 385     }
 386   /* NOTREACHED */
 387 }
 388
 389 #ifdef _LIBC
 390 /* Alias for function name in GNU C Library.  */
 391 weak_alias (__dcgettext, dcgettext);
 392 #endif
 393
 394
 395 static char *
 396 find_msg (domain_file, msgid)
 397      struct loaded_l10nfile *domain_file;
 398      const char *msgid;
 399 {
 400   size_t top, act, bottom;
 401   struct loaded_domain *domain;
 402
 403   if (domain_file->decided == 0)
 404     _nl_load_domain (domain_file);
 405
 406   if (domain_file->data == NULL)
 407     return NULL;
 408
 409   domain = (struct loaded_domain *) domain_file->data;
 410
 411   /* Locate the MSGID and its translation.  */
 412   if (domain->hash_size > 2 && domain->hash_tab != NULL)
 413     {
 414       /* Use the hashing table.  */
 415       nls_uint32 len = strlen (msgid);
 416       nls_uint32 hash_val = hash_string (msgid);
 417       nls_uint32 idx = hash_val % domain->hash_size;
 418       nls_uint32 incr = 1 + (hash_val % (domain->hash_size - 2));
 419       nls_uint32 nstr = W (domain->must_swap, domain->hash_tab[idx]);
 420
 421       if (nstr == 0)
 422         /* Hash table entry is empty.  */
 423         return NULL;
 424
 425       if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 426           && strcmp (msgid,
 427                      domain->data + W (domain->must_swap,
 428                                        domain->orig_tab[nstr - 1].offset)) == 0)
 429         return (char *) domain->data + W (domain->must_swap,
 430                                           domain->trans_tab[nstr - 1].offset);
 431
 432       while (1)
 433         {
 434           if (idx >= domain->hash_size - incr)
 435             idx -= domain->hash_size - incr;
 436           else
 437             idx += incr;
 438
 439           nstr = W (domain->must_swap, domain->hash_tab[idx]);
 440           if (nstr == 0)
 441             /* Hash table entry is empty.  */
 442             return NULL;
 443
 444           if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 445               && strcmp (msgid,
 446                          domain->data + W (domain->must_swap,
 447                                            domain->orig_tab[nstr - 1].offset))
 448                  == 0)
 449             return (char *) domain->data
 450               + W (domain->must_swap, domain->trans_tab[nstr - 1].offset);
 451         }
 452       /* NOTREACHED */
 453     }
 454
 455   /* Now we try the default method:  binary search in the sorted
 456      array of messages.  */
 457   bottom = 0;
 458   top = domain->nstrings;
 459   while (bottom < top)
 460     {
 461       int cmp_val;
 462
 463       act = (bottom + top) / 2;
 464       cmp_val = strcmp (msgid, domain->data
 465                                + W (domain->must_swap,
 466                                     domain->orig_tab[act].offset));
 467       if (cmp_val < 0)
 468         top = act;
 469       else if (cmp_val > 0)
 470         bottom = act + 1;
 471       else
 472         break;
 473     }
 474
 475   /* If an translation is found return this.  */
 476   return bottom >= top ? NULL : (char *) domain->data
 477                                 + W (domain->must_swap,
 478                                      domain->trans_tab[act].offset);
 479 }
 480
 481
 482 /* Return string representation of locale CATEGORY.  */
 483 static const char *
 484 category_to_name (category)
 485      int category;
 486 {
 487   const char *retval;
 488
 489   switch (category)
 490   {
 491 #ifdef LC_COLLATE
 492   case LC_COLLATE:
 493     retval = "LC_COLLATE";
 494     break;
 495 #endif
 496 #ifdef LC_CTYPE
 497   case LC_CTYPE:
 498     retval = "LC_CTYPE";
 499     break;
 500 #endif
 501 #ifdef LC_MONETARY
 502   case LC_MONETARY:
 503     retval = "LC_MONETARY";
 504     break;
 505 #endif
 506 #ifdef LC_NUMERIC
 507   case LC_NUMERIC:
 508     retval = "LC_NUMERIC";
 509     break;
 510 #endif
 511 #ifdef LC_TIME
 512   case LC_TIME:
 513     retval = "LC_TIME";
 514     break;
 515 #endif
 516 #ifdef LC_MESSAGES
 517   case LC_MESSAGES:
 518     retval = "LC_MESSAGES";
 519     break;
 520 #endif
 521 #ifdef LC_RESPONSE
 522   case LC_RESPONSE:
 523     retval = "LC_RESPONSE";
 524     break;
 525 #endif
 526 #ifdef LC_ALL
 527   case LC_ALL:
 528     /* This might not make sense but is perhaps better than any other
 529        value.  */
 530     retval = "LC_ALL";
 531     break;
 532 #endif
 533   default:
 534     /* If you have a better idea for a default value let me know.  */
 535     retval = "LC_XXX";
 536   }
 537
 538   return retval;
 539 }
 540
 541 /* Guess value of current locale from value of the environment variables.  */
 542 static const char *guess_category_value (category, categoryname)
 543      int category;
 544      const char *categoryname;
 545 {
 546   const char *retval;
 547
 548   /* The highest priority value is the `LANGUAGE' environment
 549      variable.  This is a GNU extension.  */
 550   retval = getenv ("LANGUAGE");
 551   if (retval != NULL && retval[0] != '\0')
 552     return retval;
 553
 554   /* `LANGUAGE' is not set.  So we have to proceed with the POSIX
 555      methods of looking to `LC_ALL', `LC_xxx', and `LANG'.  On some
 556      systems this can be done by the `setlocale' function itself.  */
 557 #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
 558   return setlocale (category, NULL);
 559 #else
 560   /* Setting of LC_ALL overwrites all other.  */
 561   retval = getenv ("LC_ALL");
 562   if (retval != NULL && retval[0] != '\0')
 563     return retval;
 564
 565   /* Next comes the name of the desired category.  */
 566   retval = getenv (categoryname);
 567   if (retval != NULL && retval[0] != '\0')
 568     return retval;
 569
 570   /* Last possibility is the LANG environment variable.  */
 571   retval = getenv ("LANG");
 572   if (retval != NULL && retval[0] != '\0')
 573     return retval;
 574
 575   /* We use C as the default domain.  POSIX says this is implementation
 576      defined.  */
 577   return "C";
 578 #endif
 579 }
 580
 581 /* @@ begin of epilog @@ */
 582
 583 /* We don't want libintl.a to depend on any other library.  So we
 584    avoid the non-standard function stpcpy.  In GNU C Library this
 585    function is available, though.  Also allow the symbol HAVE_STPCPY
 586    to be defined.  */
 587 #if !_LIBC && !HAVE_STPCPY
 588 static char *
 589 stpcpy (dest, src)
 590      char *dest;
 591      const char *src;
 592 {
 593   while ((*dest++ = *src++) != '\0')
 594     /* Do nothing. */ ;
 595   return dest - 1;
 596 }
 597 #endif