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 # ifndef _GNU_SOURCE
  63 #  define _GNU_SOURCE   1
  64 # endif
  65 # include <string.h>
  66 #else
  67 # include <strings.h>
  68 #endif
  69 #if !HAVE_STRCHR && !defined _LIBC
  70 # ifndef strchr
  71 #  define strchr index
  72 # endif
  73 #endif
  74
  75 #if defined HAVE_UNISTD_H || defined _LIBC
  76 # include <unistd.h>
  77 #endif
  78
  79 #include "gettext.h"
  80 #include "gettextP.h"
  81 #ifdef _LIBC
  82 # include <libintl.h>
  83 #else
  84 # include "libgettext.h"
  85 #endif
  86 #include "hash-string.h"
  87
  88 /* @@ end of prolog @@ */
  89
  90 #ifdef _LIBC
  91 /* Rename the non ANSI C functions.  This is required by the standard
  92    because some ANSI C functions will require linking with this object
  93    file and the name space must not be polluted.  */
  94 # define getcwd __getcwd
  95 # define stpcpy __stpcpy
  96 #else
  97 # if !defined HAVE_GETCWD
  98 char *getwd ();
  99 #  define getcwd(buf, max) getwd (buf)
 100 # else
 101 char *getcwd ();
 102 # endif
 103 # ifndef HAVE_STPCPY
 104 static char *stpcpy PARAMS ((char *dest, const char *src));
 105 # endif
 106 #endif
 107
 108 /* Amount to increase buffer size by in each try.  */
 109 #define PATH_INCR 32
 110
 111 /* The following is from pathmax.h.  */
 112 /* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define
 113    PATH_MAX but might cause redefinition warnings when sys/param.h is
 114    later included (as on MORE/BSD 4.3).  */
 115 #if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && !defined(__GNUC__))
 116 # include <limits.h>
 117 #endif
 118
 119 #ifndef _POSIX_PATH_MAX
 120 # define _POSIX_PATH_MAX 255
 121 #endif
 122
 123 #if !defined(PATH_MAX) && defined(_PC_PATH_MAX)
 124 # define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX))
 125 #endif
 126
 127 /* Don't include sys/param.h if it already has been.  */
 128 #if defined(HAVE_SYS_PARAM_H) && !defined(PATH_MAX) && !defined(MAXPATHLEN)
 129 # include <sys/param.h>
 130 #endif
 131
 132 #if !defined(PATH_MAX) && defined(MAXPATHLEN)
 133 # define PATH_MAX MAXPATHLEN
 134 #endif
 135
 136 #ifndef PATH_MAX
 137 # define PATH_MAX _POSIX_PATH_MAX
 138 #endif
 139
 140 /* XPG3 defines the result of `setlocale (category, NULL)' as:
 141    ``Directs `setlocale()' to query `category' and return the current
 142      setting of `local'.''
 143    However it does not specify the exact format.  And even worse: POSIX
 144    defines this not at all.  So we can use this feature only on selected
 145    system (e.g. those using GNU C Library).  */
 146 #ifdef _LIBC
 147 # define HAVE_LOCALE_NULL
 148 #endif
 149
 150 /* Name of the default domain used for gettext(3) prior any call to
 151    textdomain(3).  The default value for this is "messages".  */
 152 const char _nl_default_default_domain[] = "messages";
 153
 154 /* Value used as the default domain for gettext(3).  */
 155 const char *_nl_current_default_domain = _nl_default_default_domain;
 156
 157 /* Contains the default location of the message catalogs.  */
 158 const char _nl_default_dirname[] = GNULOCALEDIR;
 159
 160 /* List with bindings of specific domains created by bindtextdomain()
 161    calls.  */
 162 struct binding *_nl_domain_bindings;
 163
 164 /* Prototypes for local functions.  */
 165 static char *find_msg PARAMS ((struct loaded_l10nfile *domain_file,
 166                                const char *msgid));
 167 static const char *category_to_name PARAMS ((int category));
 168 static const char *guess_category_value PARAMS ((int category,
 169                                                  const char *categoryname));
 170
 171
 172 /* For those loosing systems which don't have `alloca' we have to add
 173    some additional code emulating it.  */
 174 #ifdef HAVE_ALLOCA
 175 /* Nothing has to be done.  */
 176 # define ADD_BLOCK(list, address) /* nothing */
 177 # define FREE_BLOCKS(list) /* nothing */
 178 #else
 179 struct block_list
 180 {
 181   void *address;
 182   struct block_list *next;
 183 };
 184 # define ADD_BLOCK(list, addr)                                                \
 185   do {                                                                        \
 186     struct block_list *newp = (struct block_list *) malloc (sizeof (*newp));  \
 187     /* If we cannot get a free block we cannot add the new element to         \
 188        the list.  */                                                          \
 189     if (newp != NULL) {                                                       \
 190       newp->address = (addr);                                                 \
 191       newp->next = (list);                                                    \
 192       (list) = newp;                                                          \
 193     }                                                                         \
 194   } while (0)
 195 # define FREE_BLOCKS(list)                                                    \
 196   do {                                                                        \
 197     while (list != NULL) {                                                    \
 198       struct block_list *old = list;                                          \
 199       list = list->next;                                                      \
 200       free (old);                                                             \
 201     }                                                                         \
 202   } while (0)
 203 # undef alloca
 204 # define alloca(size) (malloc (size))
 205 #endif  /* have alloca */
 206
 207
 208 /* Names for the libintl functions are a problem.  They must not clash
 209    with existing names and they should follow ANSI C.  But this source
 210    code is also used in GNU C Library where the names have a __
 211    prefix.  So we have to make a difference here.  */
 212 #ifdef _LIBC
 213 # define DCGETTEXT __dcgettext
 214 #else
 215 # define DCGETTEXT dcgettext__
 216 #endif
 217
 218 /* Look up MSGID in the DOMAINNAME message catalog for the current CATEGORY
 219    locale.  */
 220 char *
 221 DCGETTEXT (domainname, msgid, category)
 222      const char *domainname;
 223      const char *msgid;
 224      int category;
 225 {
 226 #ifndef HAVE_ALLOCA
 227   struct block_list *block_list = NULL;
 228 #endif
 229   struct loaded_l10nfile *domain;
 230   struct binding *binding;
 231   const char *categoryname;
 232   const char *categoryvalue;
 233   char *dirname, *xdomainname;
 234   char *single_locale;
 235   char *retval;
 236   int saved_errno = errno;
 237
 238   /* If no real MSGID is given return NULL.  */
 239   if (msgid == NULL)
 240     return NULL;
 241
 242   /* If DOMAINNAME is NULL, we are interested in the default domain.  If
 243      CATEGORY is not LC_MESSAGES this might not make much sense but the
 244      defintion left this undefined.  */
 245   if (domainname == NULL)
 246     domainname = _nl_current_default_domain;
 247
 248   /* First find matching binding.  */
 249   for (binding = _nl_domain_bindings; binding != NULL; binding = binding->next)
 250     {
 251       int compare = strcmp (domainname, binding->domainname);
 252       if (compare == 0)
 253         /* We found it!  */
 254         break;
 255       if (compare < 0)
 256         {
 257           /* It is not in the list.  */
 258           binding = NULL;
 259           break;
 260         }
 261     }
 262
 263   if (binding == NULL)
 264     dirname = (char *) _nl_default_dirname;
 265   else if (binding->dirname[0] == '/')
 266     dirname = binding->dirname;
 267   else
 268     {
 269       /* We have a relative path.  Make it absolute now.  */
 270       size_t dirname_len = strlen (binding->dirname) + 1;
 271       size_t path_max;
 272       char *ret;
 273
 274       path_max = (unsigned) PATH_MAX;
 275       path_max += 2;            /* The getcwd docs say to do this.  */
 276
 277       dirname = (char *) alloca (path_max + dirname_len);
 278       ADD_BLOCK (block_list, dirname);
 279
 280       errno = 0;
 281       while ((ret = getcwd (dirname, path_max)) == NULL && errno == ERANGE)
 282         {
 283           path_max += PATH_INCR;
 284           dirname = (char *) alloca (path_max + dirname_len);
 285           ADD_BLOCK (block_list, dirname);
 286           errno = 0;
 287         }
 288
 289       if (ret == NULL)
 290         {
 291           /* We cannot get the current working directory.  Don't signal an
 292              error but simply return the default string.  */
 293           FREE_BLOCKS (block_list);
 294           errno = saved_errno;
 295           return (char *) msgid;
 296         }
 297
 298       /* We don't want libintl.a to depend on any other library.  So
 299          we avoid the non-standard function stpcpy.  In GNU C Library
 300          this function is available, though.  Also allow the symbol
 301          HAVE_STPCPY to be defined.  */
 302       stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
 303     }
 304
 305   /* Now determine the symbolic name of CATEGORY and its value.  */
 306   categoryname = category_to_name (category);
 307   categoryvalue = guess_category_value (category, categoryname);
 308
 309   xdomainname = (char *) alloca (strlen (categoryname)
 310                                  + strlen (domainname) + 5);
 311   ADD_BLOCK (block_list, xdomainname);
 312   /* We don't want libintl.a to depend on any other library.  So we
 313      avoid the non-standard function stpcpy.  In GNU C Library this
 314      function is available, though.  Also allow the symbol HAVE_STPCPY
 315      to be defined.  */
 316   stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
 317                   domainname),
 318           ".mo");
 319
 320   /* Creating working area.  */
 321   single_locale = (char *) alloca (strlen (categoryvalue) + 1);
 322   ADD_BLOCK (block_list, single_locale);
 323
 324
 325   /* Search for the given string.  This is a loop because we perhaps
 326      got an ordered list of languages to consider for th translation.  */
 327   while (1)
 328     {
 329       /* Make CATEGORYVALUE point to the next element of the list.  */
 330       while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
 331         ++categoryvalue;
 332       if (categoryvalue[0] == '\0')
 333         {
 334           /* The whole contents of CATEGORYVALUE has been searched but
 335              no valid entry has been found.  We solve this situation
 336              by implicitely appending a "C" entry, i.e. no translation
 337              will take place.  */
 338           single_locale[0] = 'C';
 339           single_locale[1] = '\0';
 340         }
 341       else
 342         {
 343           char *cp = single_locale;
 344           while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
 345             *cp++ = *categoryvalue++;
 346           *cp = '\0';
 347         }
 348
 349       /* If the current locale value is C (or POSIX) we don't load a
 350          domain.  Return the MSGID.  */
 351       if (strcmp (single_locale, "C") == 0
 352           || strcmp (single_locale, "POSIX") == 0)
 353         {
 354           FREE_BLOCKS (block_list);
 355           errno = saved_errno;
 356           return (char *) msgid;
 357         }
 358
 359
 360       /* Find structure describing the message catalog matching the
 361          DOMAINNAME and CATEGORY.  */
 362       domain = _nl_find_domain (dirname, single_locale, xdomainname);
 363
 364       if (domain != NULL)
 365         {
 366           retval = find_msg (domain, msgid);
 367
 368           if (retval == NULL)
 369             {
 370               int cnt;
 371
 372               for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
 373                 {
 374                   retval = find_msg (domain->successor[cnt], msgid);
 375
 376                   if (retval != NULL)
 377                     break;
 378                 }
 379             }
 380
 381           if (retval != NULL)
 382             {
 383               FREE_BLOCKS (block_list);
 384               errno = saved_errno;
 385               return retval;
 386             }
 387         }
 388     }
 389   /* NOTREACHED */
 390 }
 391
 392 #ifdef _LIBC
 393 /* Alias for function name in GNU C Library.  */
 394 weak_alias (__dcgettext, dcgettext);
 395 #endif
 396
 397
 398 static char *
 399 find_msg (domain_file, msgid)
 400      struct loaded_l10nfile *domain_file;
 401      const char *msgid;
 402 {
 403   size_t top, act, bottom;
 404   struct loaded_domain *domain;
 405
 406   if (domain_file->decided == 0)
 407     _nl_load_domain (domain_file);
 408
 409   if (domain_file->data == NULL)
 410     return NULL;
 411
 412   domain = (struct loaded_domain *) domain_file->data;
 413
 414   /* Locate the MSGID and its translation.  */
 415   if (domain->hash_size > 2 && domain->hash_tab != NULL)
 416     {
 417       /* Use the hashing table.  */
 418       nls_uint32 len = strlen (msgid);
 419       nls_uint32 hash_val = hash_string (msgid);
 420       nls_uint32 idx = hash_val % domain->hash_size;
 421       nls_uint32 incr = 1 + (hash_val % (domain->hash_size - 2));
 422       nls_uint32 nstr = W (domain->must_swap, domain->hash_tab[idx]);
 423
 424       if (nstr == 0)
 425         /* Hash table entry is empty.  */
 426         return NULL;
 427
 428       if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 429           && strcmp (msgid,
 430                      domain->data + W (domain->must_swap,
 431                                        domain->orig_tab[nstr - 1].offset)) == 0)
 432         return (char *) domain->data + W (domain->must_swap,
 433                                           domain->trans_tab[nstr - 1].offset);
 434
 435       while (1)
 436         {
 437           if (idx >= domain->hash_size - incr)
 438             idx -= domain->hash_size - incr;
 439           else
 440             idx += incr;
 441
 442           nstr = W (domain->must_swap, domain->hash_tab[idx]);
 443           if (nstr == 0)
 444             /* Hash table entry is empty.  */
 445             return NULL;
 446
 447           if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 448               && strcmp (msgid,
 449                          domain->data + W (domain->must_swap,
 450                                            domain->orig_tab[nstr - 1].offset))
 451                  == 0)
 452             return (char *) domain->data
 453               + W (domain->must_swap, domain->trans_tab[nstr - 1].offset);
 454         }
 455       /* NOTREACHED */
 456     }
 457
 458   /* Now we try the default method:  binary search in the sorted
 459      array of messages.  */
 460   bottom = 0;
 461   top = domain->nstrings;
 462   while (bottom < top)
 463     {
 464       int cmp_val;
 465
 466       act = (bottom + top) / 2;
 467       cmp_val = strcmp (msgid, domain->data
 468                                + W (domain->must_swap,
 469                                     domain->orig_tab[act].offset));
 470       if (cmp_val < 0)
 471         top = act;
 472       else if (cmp_val > 0)
 473         bottom = act + 1;
 474       else
 475         break;
 476     }
 477
 478   /* If an translation is found return this.  */
 479   return bottom >= top ? NULL : (char *) domain->data
 480                                 + W (domain->must_swap,
 481                                      domain->trans_tab[act].offset);
 482 }
 483
 484
 485 /* Return string representation of locale CATEGORY.  */
 486 static const char *
 487 category_to_name (category)
 488      int category;
 489 {
 490   const char *retval;
 491
 492   switch (category)
 493   {
 494 #ifdef LC_COLLATE
 495   case LC_COLLATE:
 496     retval = "LC_COLLATE";
 497     break;
 498 #endif
 499 #ifdef LC_CTYPE
 500   case LC_CTYPE:
 501     retval = "LC_CTYPE";
 502     break;
 503 #endif
 504 #ifdef LC_MONETARY
 505   case LC_MONETARY:
 506     retval = "LC_MONETARY";
 507     break;
 508 #endif
 509 #ifdef LC_NUMERIC
 510   case LC_NUMERIC:
 511     retval = "LC_NUMERIC";
 512     break;
 513 #endif
 514 #ifdef LC_TIME
 515   case LC_TIME:
 516     retval = "LC_TIME";
 517     break;
 518 #endif
 519 #ifdef LC_MESSAGES
 520   case LC_MESSAGES:
 521     retval = "LC_MESSAGES";
 522     break;
 523 #endif
 524 #ifdef LC_RESPONSE
 525   case LC_RESPONSE:
 526     retval = "LC_RESPONSE";
 527     break;
 528 #endif
 529 #ifdef LC_ALL
 530   case LC_ALL:
 531     /* This might not make sense but is perhaps better than any other
 532        value.  */
 533     retval = "LC_ALL";
 534     break;
 535 #endif
 536   default:
 537     /* If you have a better idea for a default value let me know.  */
 538     retval = "LC_XXX";
 539   }
 540
 541   return retval;
 542 }
 543
 544 /* Guess value of current locale from value of the environment variables.  */
 545 static const char *guess_category_value (category, categoryname)
 546      int category;
 547      const char *categoryname;
 548 {
 549   const char *retval;
 550
 551   /* The highest priority value is the `LANGUAGE' environment
 552      variable.  This is a GNU extension.  */
 553   retval = getenv ("LANGUAGE");
 554   if (retval != NULL && retval[0] != '\0')
 555     return retval;
 556
 557   /* `LANGUAGE' is not set.  So we have to proceed with the POSIX
 558      methods of looking to `LC_ALL', `LC_xxx', and `LANG'.  On some
 559      systems this can be done by the `setlocale' function itself.  */
 560 #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
 561   return setlocale (category, NULL);
 562 #else
 563   /* Setting of LC_ALL overwrites all other.  */
 564   retval = getenv ("LC_ALL");
 565   if (retval != NULL && retval[0] != '\0')
 566     return retval;
 567
 568   /* Next comes the name of the desired category.  */
 569   retval = getenv (categoryname);
 570   if (retval != NULL && retval[0] != '\0')
 571     return retval;
 572
 573   /* Last possibility is the LANG environment variable.  */
 574   retval = getenv ("LANG");
 575   if (retval != NULL && retval[0] != '\0')
 576     return retval;
 577
 578   /* We use C as the default domain.  POSIX says this is implementation
 579      defined.  */
 580   return "C";
 581 #endif
 582 }
 583
 584 /* @@ begin of epilog @@ */
 585
 586 /* We don't want libintl.a to depend on any other library.  So we
 587    avoid the non-standard function stpcpy.  In GNU C Library this
 588    function is available, though.  Also allow the symbol HAVE_STPCPY
 589    to be defined.  */
 590 #if !_LIBC && !HAVE_STPCPY
 591 static char *
 592 stpcpy (dest, src)
 593      char *dest;
 594      const char *src;
 595 {
 596   while ((*dest++ = *src++) != '\0')
 597     /* Do nothing. */ ;
 598   return dest - 1;
 599 }
 600 #endif