intl/dcgettext.c

   1 /* dcgettext.c -- implementation of the dcgettext(3) function
   2    Copyright (C) 1995 Free Software Foundation, Inc.
   3
   4 This program is free software; you can redistribute it and/or modify
   5 it under the terms of the GNU General Public License as published by
   6 the Free Software Foundation; either version 2, or (at your option)
   7 any later version.
   8
   9 This program is distributed in the hope that it will be useful,
  10 but WITHOUT ANY WARRANTY; without even the implied warranty of
  11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12 GNU General Public License for more details.
  13
  14 You should have received a copy of the GNU General Public License
  15 along with this program; if not, write to the Free Software
  16 Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  17
  18 #ifdef HAVE_CONFIG_H
  19 # include <config.h>
  20 #endif
  21
  22 #include <sys/types.h>
  23
  24 #ifdef __GNUC__
  25 # define alloca __builtin_alloca
  26 #else
  27 # if defined HAVE_ALLOCA_H || defined _LIBC
  28 #  include <alloca.h>
  29 # else
  30 #  ifdef _AIX
  31  #pragma alloca
  32 #  else
  33 #   ifndef alloca
  34 char *alloca ();
  35 #   endif
  36 #  endif
  37 # endif
  38 #endif
  39
  40 #include <errno.h>
  41 #ifndef errno
  42 extern int errno;
  43 #endif
  44
  45 #if defined STDC_HEADERS || defined _LIBC
  46 # include <stdlib.h>
  47 #else
  48 char *getenv ();
  49 # ifdef HAVE_MALLOC_H
  50 #  include <malloc.h>
  51 # else
  52 void free ();
  53 # endif
  54 #endif
  55
  56 #if defined HAVE_STRING_H || defined _LIBC
  57 # include <string.h>
  58 #else
  59 # include <strings.h>
  60 #endif
  61 #if !HAVE_STRCHR && !defined _LIBC
  62 # ifndef strchr
  63 #  define strchr index
  64 # endif
  65 #endif
  66
  67 #if defined HAVE_UNISTD_H || defined _LIBC
  68 # include <unistd.h>
  69 #endif
  70
  71 #include "gettext.h"
  72 #include "gettextP.h"
  73 #ifdef _LIBC
  74 # include <libintl.h>
  75 #else
  76 # include "libgettext.h"
  77 #endif
  78 #include "hash-string.h"
  79
  80 /* @@ end of prolog @@ */
  81
  82 #ifdef _LIBC
  83 /* Rename the non ANSI C functions.  This is required by the standard
  84    because some ANSI C functions will require linking with this object
  85    file and the name space must not be polluted.  */
  86 # define getcwd __getcwd
  87 # define stpcpy __stpcpy
  88 #endif
  89
  90 #if !defined HAVE_GETCWD && !defined _LIBC
  91 char *getwd ();
  92 # define getcwd(buf, max) getwd (buf)
  93 #else
  94 char *getcwd ();
  95 #endif
  96
  97 /* Amount to increase buffer size by in each try.  */
  98 #define PATH_INCR 32
  99
 100 /* The following is from pathmax.h.  */
 101 /* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define
 102    PATH_MAX but might cause redefinition warnings when sys/param.h is
 103    later included (as on MORE/BSD 4.3).  */
 104 #if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && !defined(__GNUC__))
 105 # include <limits.h>
 106 #endif
 107
 108 #ifndef _POSIX_PATH_MAX
 109 # define _POSIX_PATH_MAX 255
 110 #endif
 111
 112 #if !defined(PATH_MAX) && defined(_PC_PATH_MAX)
 113 # define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX))
 114 #endif
 115
 116 /* Don't include sys/param.h if it already has been.  */
 117 #if defined(HAVE_SYS_PARAM_H) && !defined(PATH_MAX) && !defined(MAXPATHLEN)
 118 # include <sys/param.h>
 119 #endif
 120
 121 #if !defined(PATH_MAX) && defined(MAXPATHLEN)
 122 # define PATH_MAX MAXPATHLEN
 123 #endif
 124
 125 #ifndef PATH_MAX
 126 # define PATH_MAX _POSIX_PATH_MAX
 127 #endif
 128
 129 /* XPG3 defines the result of `setlocale (category, NULL)' as:
 130    ``Directs `setlocale()' to query `category' and return the current
 131      setting of `local'.''
 132    However it does not specify the exact format.  And even worse: POSIX
 133    defines this not at all.  So we can use this feature only on selected
 134    system (e.g. those using GNU C Library).  */
 135 #ifdef _LIBC
 136 # define HAVE_LOCALE_NULL
 137 #endif
 138
 139 /* Name of the default domain used for gettext(3) prior any call to
 140    textdomain(3).  The default value for this is "messages".  */
 141 const char _nl_default_default_domain[] = "messages";
 142
 143 /* Value used as the default domain for gettext(3).  */
 144 const char *_nl_current_default_domain = _nl_default_default_domain;
 145
 146 /* Contains the default location of the message catalogs.  */
 147 const char _nl_default_dirname[] = GNULOCALEDIR;
 148
 149 /* List with bindings of specific domains created by bindtextdomain()
 150    calls.  */
 151 struct binding *_nl_domain_bindings;
 152
 153 /* Prototypes for local functions.  */
 154 static char *find_msg PARAMS ((struct loaded_domain *domain,
 155                                const char *msgid));
 156 static const char *category_to_name PARAMS ((int category));
 157 static const char *guess_category_value PARAMS ((int category,
 158                                                  const char *categoryname));
 159
 160
 161 /* Names for the libintl functions are a problem.  They must not clash
 162    with existing names and they should follow ANSI C.  But this source
 163    code is also used in GNU C Library where the names have a __
 164    prefix.  So we have to make a difference here.  */
 165 #ifdef _LIBC
 166 # define DCGETTEXT __dcgettext
 167 #else
 168 # define DCGETTEXT dcgettext__
 169 #endif
 170
 171 /* Look up MSGID in the DOMAINNAME message catalog for the current CATEGORY
 172    locale.  */
 173 char *
 174 DCGETTEXT (domainname, msgid, category)
 175      const char *domainname;
 176      const char *msgid;
 177      int category;
 178 {
 179   struct loaded_domain *domain;
 180   struct binding *binding;
 181   const char *categoryname;
 182   const char *categoryvalue;
 183   char *dirname, *xdomainname;
 184   char *single_locale;
 185   char *retval;
 186   int saved_errno = errno;
 187
 188   /* If no real MSGID is given return NULL.  */
 189   if (msgid == NULL)
 190     return NULL;
 191
 192   /* If DOMAINNAME is NULL, we are interested in the default domain.  If
 193      CATEGORY is not LC_MESSAGES this might not make much sense but the
 194      defintion left this undefined.  */
 195   if (domainname == NULL)
 196     domainname = _nl_current_default_domain;
 197
 198   /* First find matching binding.  */
 199   for (binding = _nl_domain_bindings; binding != NULL; binding = binding->next)
 200     {
 201       int compare = strcmp (domainname, binding->domainname);
 202       if (compare == 0)
 203         /* We found it!  */
 204         break;
 205       if (compare < 0)
 206         {
 207           /* It is not in the list.  */
 208           binding = NULL;
 209           break;
 210         }
 211     }
 212
 213   if (binding == NULL)
 214     dirname = (char *) _nl_default_dirname;
 215   else if (binding->dirname[0] == '/')
 216     dirname = binding->dirname;
 217   else
 218     {
 219       /* We have a relative path.  Make it absolute now.  */
 220       size_t dirname_len = strlen (binding->dirname) + 1;
 221       size_t path_max;
 222       char *ret;
 223
 224       path_max = (unsigned) PATH_MAX;
 225       path_max += 2;            /* The getcwd docs say to do this.  */
 226
 227       dirname = (char *) alloca (path_max + dirname_len);
 228
 229       errno = 0;
 230       while ((ret = getcwd (dirname, path_max)) == NULL && errno == ERANGE)
 231         {
 232           path_max += PATH_INCR;
 233           dirname = (char *) alloca (path_max + dirname_len);
 234           errno = 0;
 235         }
 236
 237       if (ret == NULL)
 238         {
 239           /* We cannot get the current working directory.  Don't signal an
 240              error but simply return the default string.  */
 241           errno = saved_errno;
 242           return (char *) msgid;
 243         }
 244
 245       /* We don't want libintl.a to depend on any other library.  So
 246          we avoid the non-standard function stpcpy.  In GNU C Library
 247          this function is available, though.  Also allow the symbol
 248          HAVE_STPCPY to be defined.  */
 249 #if defined _LIBC || defined HAVE_STPCPY
 250       stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
 251 #else
 252       strcat (dirname, "/");
 253       strcat (dirname, binding->dirname);
 254 #endif
 255     }
 256
 257   /* Now determine the symbolic name of CATEGORY and its value.  */
 258   categoryname = category_to_name (category);
 259   categoryvalue = guess_category_value (category, categoryname);
 260
 261   xdomainname = (char *) alloca (strlen (categoryname)
 262                                  + strlen (domainname) + 5);
 263   /* We don't want libintl.a to depend on any other library.  So we
 264      avoid the non-standard function stpcpy.  In GNU C Library this
 265      function is available, though.  Also allow the symbol HAVE_STPCPY
 266      to be defined.  */
 267 #if defined _LIBC || defined HAVE_STPCPY
 268   stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
 269                   domainname),
 270           ".mo");
 271 #else
 272   strcpy (xdomainname, categoryname);
 273   strcat (xdomainname, "/");
 274   strcat (xdomainname, domainname);
 275   strcat (xdomainname, ".mo");
 276 #endif
 277
 278   /* Creating working area.  */
 279   single_locale = (char *) alloca (strlen (categoryvalue) + 1);
 280
 281
 282   /* Search for the given string.  This is a loop because we perhaps
 283      got an ordered list of languages to consider for th translation.  */
 284   while (1)
 285     {
 286       /* Make CATEGORYVALUE point to the next element of the list.  */
 287       while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
 288         ++categoryvalue;
 289       if (categoryvalue[0] == '\0')
 290         {
 291           /* The whole contents of CATEGORYVALUE has been searched but
 292              no valid entry has been found.  We solve this situation
 293              by implicitely appending a "C" entry, i.e. no translation
 294              will take place.  */
 295           single_locale[0] = 'C';
 296           single_locale[1] = '\0';
 297         }
 298       else
 299         {
 300           char *cp = single_locale;
 301           while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
 302             *cp++ = *categoryvalue++;
 303           *cp = '\0';
 304         }
 305
 306       /* If the current locale value is C (or POSIX) we don't load a
 307          domain.  Return the MSGID.  */
 308       if (strcmp (single_locale, "C") == 0
 309           || strcmp (single_locale, "POSIX") == 0)
 310         {
 311           errno = saved_errno;
 312           return (char *) msgid;
 313         }
 314
 315
 316       /* Find structure describing the message catalog matching the
 317          DOMAINNAME and CATEGORY.  */
 318       domain = _nl_find_domain (dirname, single_locale, xdomainname);
 319
 320       if (domain != NULL)
 321         {
 322           retval = find_msg (domain, msgid);
 323
 324           if (retval == NULL)
 325             {
 326               int cnt;
 327
 328               for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
 329                 {
 330                   retval = find_msg (domain->successor[cnt], msgid);
 331
 332                   if (retval != NULL)
 333                     break;
 334                 }
 335             }
 336
 337           if (retval != NULL)
 338             {
 339               errno = saved_errno;
 340               return retval;
 341             }
 342         }
 343     }
 344   /* NOTREACHED */
 345 }
 346
 347 #ifdef _LIBC
 348 /* Alias for function name in GNU C Library.  */
 349 weak_alias (__dcgettext, dcgettext);
 350 #endif
 351
 352
 353 static char *
 354 find_msg (domain, msgid)
 355      struct loaded_domain *domain;
 356      const char *msgid;
 357 {
 358   size_t top, act, bottom;
 359
 360   if (domain->decided == 0)
 361     _nl_load_domain (domain);
 362
 363   if (domain->data == NULL)
 364     return NULL;
 365
 366   /* Locate the MSGID and its translation.  */
 367   if (domain->hash_size > 2 && domain->hash_tab != NULL)
 368     {
 369       /* Use the hashing table.  */
 370       nls_uint32 len = strlen (msgid);
 371       nls_uint32 hash_val = hash_string (msgid);
 372       nls_uint32 idx = hash_val % domain->hash_size;
 373       nls_uint32 incr = 1 + (hash_val % (domain->hash_size - 2));
 374       nls_uint32 nstr = W (domain->must_swap, domain->hash_tab[idx]);
 375
 376       if (nstr == 0)
 377         /* Hash table entry is empty.  */
 378         return NULL;
 379
 380       if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 381           && strcmp (msgid,
 382                      domain->data + W (domain->must_swap,
 383                                        domain->orig_tab[nstr - 1].offset)) == 0)
 384         return (char *) domain->data + W (domain->must_swap,
 385                                           domain->trans_tab[nstr - 1].offset);
 386
 387       while (1)
 388         {
 389           if (idx >= domain->hash_size - incr)
 390             idx -= domain->hash_size - incr;
 391           else
 392             idx += incr;
 393
 394           nstr = W (domain->must_swap, domain->hash_tab[idx]);
 395           if (nstr == 0)
 396             /* Hash table entry is empty.  */
 397             return NULL;
 398
 399           if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 400               && strcmp (msgid,
 401                          domain->data + W (domain->must_swap,
 402                                            domain->orig_tab[nstr - 1].offset))
 403                  == 0)
 404             return (char *) domain->data
 405               + W (domain->must_swap, domain->trans_tab[nstr - 1].offset);
 406         }
 407       /* NOTREACHED */
 408     }
 409
 410   /* Now we try the default method:  binary search in the sorted
 411      array of messages.  */
 412   bottom = 0;
 413   top = domain->nstrings;
 414   while (bottom < top)
 415     {
 416       int cmp_val;
 417
 418       act = (bottom + top) / 2;
 419       cmp_val = strcmp (msgid, domain->data
 420                                + W (domain->must_swap,
 421                                     domain->orig_tab[act].offset));
 422       if (cmp_val < 0)
 423         top = act;
 424       else if (cmp_val > 0)
 425         bottom = act + 1;
 426       else
 427         break;
 428     }
 429
 430   /* If an translation is found return this.  */
 431   return bottom >= top ? NULL : (char *) domain->data
 432                                 + W (domain->must_swap,
 433                                      domain->trans_tab[act].offset);
 434 }
 435
 436
 437 /* Return string representation of locale CATEGORY.  */
 438 static const char *category_to_name (category)
 439      int category;
 440 {
 441   const char *retval;
 442
 443   switch (category)
 444   {
 445 #ifdef LC_COLLATE
 446   case LC_COLLATE:
 447     retval = "LC_COLLATE";
 448     break;
 449 #endif
 450 #ifdef LC_CTYPE
 451   case LC_CTYPE:
 452     retval = "LC_CTYPE";
 453     break;
 454 #endif
 455 #ifdef LC_MONETARY
 456   case LC_MONETARY:
 457     retval = "LC_MONETARY";
 458     break;
 459 #endif
 460 #ifdef LC_NUMERIC
 461   case LC_NUMERIC:
 462     retval = "LC_NUMERIC";
 463     break;
 464 #endif
 465 #ifdef LC_TIME
 466   case LC_TIME:
 467     retval = "LC_TIME";
 468     break;
 469 #endif
 470 #ifdef LC_MESSAGES
 471   case LC_MESSAGES:
 472     retval = "LC_MESSAGES";
 473     break;
 474 #endif
 475 #ifdef LC_RESPONSE
 476   case LC_RESPONSE:
 477     retval = "LC_RESPONSE";
 478     break;
 479 #endif
 480 #ifdef LC_ALL
 481   case LC_ALL:
 482     /* This might not make sense but is perhaps better than any other
 483        value.  */
 484     retval = "LC_ALL";
 485     break;
 486 #endif
 487   default:
 488     /* If you have a better idea for a default value let me know.  */
 489     retval = "LC_XXX";
 490   }
 491
 492   return retval;
 493 }
 494
 495 /* Guess value of current locale from value of the environment variables.  */
 496 static const char *guess_category_value (category, categoryname)
 497      int category;
 498      const char *categoryname;
 499 {
 500   const char *retval;
 501
 502   /* The highest priority value is the `LANGUAGE' environment
 503      variable.  This is a GNU extension.  */
 504   retval = getenv ("LANGUAGE");
 505   if (retval != NULL && retval[0] != '\0')
 506     return retval;
 507
 508   /* `LANGUAGE' is not set.  So we have to proceed with the POSIX
 509      methods of looking to `LC_ALL', `LC_xxx', and `LANG'.  On some
 510      systems this can be done by the `setlocale' function itself.  */
 511 #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
 512   return setlocale (category, NULL);
 513 #else
 514   /* Setting of LC_ALL overwrites all other.  */
 515   retval = getenv ("LC_ALL");
 516   if (retval != NULL && retval[0] != '\0')
 517     return retval;
 518
 519   /* Next comes the name of the desired category.  */
 520   retval = getenv (categoryname);
 521   if (retval != NULL && retval[0] != '\0')
 522     return retval;
 523
 524   /* Last possibility is the LANG environment variable.  */
 525   retval = getenv ("LANG");
 526   if (retval != NULL && retval[0] != '\0')
 527     return retval;
 528
 529   /* We use C as the default domain.  POSIX says this is implementation
 530      defined.  */
 531   return "C";
 532 #endif
 533 }