intl/dcgettext.c

   1 /* Implementation of the dcgettext(3) function.
   2    Copyright (C) 1995, 1996, 1997, 1998 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 Foundation,
  16    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 # define HAVE_ALLOCA 1
  27 #else
  28 # if defined HAVE_ALLOCA_H || defined _LIBC
  29 #  include <alloca.h>
  30 # else
  31 #  ifdef _AIX
  32  #pragma alloca
  33 #  else
  34 #   ifndef alloca
  35 char *alloca ();
  36 #   endif
  37 #  endif
  38 # endif
  39 #endif
  40
  41 #include <errno.h>
  42 #ifndef errno
  43 extern int errno;
  44 #endif
  45 #ifndef __set_errno
  46 # define __set_errno(val) errno = (val)
  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 # ifndef _GNU_SOURCE
  62 #  define _GNU_SOURCE   1
  63 # endif
  64 # include <string.h>
  65 #else
  66 # include <strings.h>
  67 #endif
  68 #if !HAVE_STRCHR && !defined _LIBC
  69 # ifndef strchr
  70 #  define strchr index
  71 # endif
  72 #endif
  73
  74 #if defined HAVE_UNISTD_H || defined _LIBC
  75 # include <unistd.h>
  76 #endif
  77
  78 #include "gettext.h"
  79 #include "gettextP.h"
  80 #ifdef _LIBC
  81 # include <libintl.h>
  82 #else
  83 # include "libgettext.h"
  84 #endif
  85 #include "hash-string.h"
  86
  87 /* @@ end of prolog @@ */
  88
  89 #ifdef _LIBC
  90 /* Rename the non ANSI C functions.  This is required by the standard
  91    because some ANSI C functions will require linking with this object
  92    file and the name space must not be polluted.  */
  93 # define getcwd __getcwd
  94 # ifndef stpcpy
  95 #  define stpcpy __stpcpy
  96 # endif
  97 #else
  98 # if !defined HAVE_GETCWD
  99 char *getwd ();
 100 #  define getcwd(buf, max) getwd (buf)
 101 # else
 102 char *getcwd ();
 103 # endif
 104 # ifndef HAVE_STPCPY
 105 static char *stpcpy PARAMS ((char *dest, const char *src));
 106 # endif
 107 #endif
 108
 109 /* Amount to increase buffer size by in each try.  */
 110 #define PATH_INCR 32
 111
 112 /* The following is from pathmax.h.  */
 113 /* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define
 114    PATH_MAX but might cause redefinition warnings when sys/param.h is
 115    later included (as on MORE/BSD 4.3).  */
 116 #if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && !defined(__GNUC__))
 117 # include <limits.h>
 118 #endif
 119
 120 #ifndef _POSIX_PATH_MAX
 121 # define _POSIX_PATH_MAX 255
 122 #endif
 123
 124 #if !defined(PATH_MAX) && defined(_PC_PATH_MAX)
 125 # define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX))
 126 #endif
 127
 128 /* Don't include sys/param.h if it already has been.  */
 129 #if defined(HAVE_SYS_PARAM_H) && !defined(PATH_MAX) && !defined(MAXPATHLEN)
 130 # include <sys/param.h>
 131 #endif
 132
 133 #if !defined(PATH_MAX) && defined(MAXPATHLEN)
 134 # define PATH_MAX MAXPATHLEN
 135 #endif
 136
 137 #ifndef PATH_MAX
 138 # define PATH_MAX _POSIX_PATH_MAX
 139 #endif
 140
 141 /* XPG3 defines the result of `setlocale (category, NULL)' as:
 142    ``Directs `setlocale()' to query `category' and return the current
 143      setting of `local'.''
 144    However it does not specify the exact format.  And even worse: POSIX
 145    defines this not at all.  So we can use this feature only on selected
 146    system (e.g. those using GNU C Library).  */
 147 #ifdef _LIBC
 148 # define HAVE_LOCALE_NULL
 149 #endif
 150
 151 /* Name of the default domain used for gettext(3) prior any call to
 152    textdomain(3).  The default value for this is "messages".  */
 153 const char _nl_default_default_domain[] = "messages";
 154
 155 /* Value used as the default domain for gettext(3).  */
 156 const char *_nl_current_default_domain = _nl_default_default_domain;
 157
 158 /* Contains the default location of the message catalogs.  */
 159 const char _nl_default_dirname[] = GNULOCALEDIR;
 160
 161 /* List with bindings of specific domains created by bindtextdomain()
 162    calls.  */
 163 struct binding *_nl_domain_bindings;
 164
 165 /* Prototypes for local functions.  */
 166 static char *find_msg PARAMS ((struct loaded_l10nfile *domain_file,
 167                                const char *msgid)) internal_function;
 168 static const char *category_to_name PARAMS ((int category)) internal_function;
 169 static const char *guess_category_value PARAMS ((int category,
 170                                                  const char *categoryname))
 171      internal_function;
 172
 173
 174 /* For those loosing systems which don't have `alloca' we have to add
 175    some additional code emulating it.  */
 176 #ifdef HAVE_ALLOCA
 177 /* Nothing has to be done.  */
 178 # define ADD_BLOCK(list, address) /* nothing */
 179 # define FREE_BLOCKS(list) /* nothing */
 180 #else
 181 struct block_list
 182 {
 183   void *address;
 184   struct block_list *next;
 185 };
 186 # define ADD_BLOCK(list, addr)                                                \
 187   do {                                                                        \
 188     struct block_list *newp = (struct block_list *) malloc (sizeof (*newp));  \
 189     /* If we cannot get a free block we cannot add the new element to         \
 190        the list.  */                                                          \
 191     if (newp != NULL) {                                                       \
 192       newp->address = (addr);                                                 \
 193       newp->next = (list);                                                    \
 194       (list) = newp;                                                          \
 195     }                                                                         \
 196   } while (0)
 197 # define FREE_BLOCKS(list)                                                    \
 198   do {                                                                        \
 199     while (list != NULL) {                                                    \
 200       struct block_list *old = list;                                          \
 201       list = list->next;                                                      \
 202       free (old);                                                             \
 203     }                                                                         \
 204   } while (0)
 205 # undef alloca
 206 # define alloca(size) (malloc (size))
 207 #endif  /* have alloca */
 208
 209
 210 /* Names for the libintl functions are a problem.  They must not clash
 211    with existing names and they should follow ANSI C.  But this source
 212    code is also used in GNU C Library where the names have a __
 213    prefix.  So we have to make a difference here.  */
 214 #ifdef _LIBC
 215 # define DCGETTEXT __dcgettext
 216 #else
 217 # define DCGETTEXT dcgettext__
 218 #endif
 219
 220 /* Look up MSGID in the DOMAINNAME message catalog for the current CATEGORY
 221    locale.  */
 222 char *
 223 DCGETTEXT (domainname, msgid, category)
 224      const char *domainname;
 225      const char *msgid;
 226      int category;
 227 {
 228 #ifndef HAVE_ALLOCA
 229   struct block_list *block_list = NULL;
 230 #endif
 231   struct loaded_l10nfile *domain;
 232   struct binding *binding;
 233   const char *categoryname;
 234   const char *categoryvalue;
 235   char *dirname, *xdomainname;
 236   char *single_locale;
 237   char *retval;
 238   int saved_errno = errno;
 239
 240   /* If no real MSGID is given return NULL.  */
 241   if (msgid == NULL)
 242     return NULL;
 243
 244   /* If DOMAINNAME is NULL, we are interested in the default domain.  If
 245      CATEGORY is not LC_MESSAGES this might not make much sense but the
 246      defintion left this undefined.  */
 247   if (domainname == NULL)
 248     domainname = _nl_current_default_domain;
 249
 250   /* First find matching binding.  */
 251   for (binding = _nl_domain_bindings; binding != NULL; binding = binding->next)
 252     {
 253       int compare = strcmp (domainname, binding->domainname);
 254       if (compare == 0)
 255         /* We found it!  */
 256         break;
 257       if (compare < 0)
 258         {
 259           /* It is not in the list.  */
 260           binding = NULL;
 261           break;
 262         }
 263     }
 264
 265   if (binding == NULL)
 266     dirname = (char *) _nl_default_dirname;
 267   else if (binding->dirname[0] == '/')
 268     dirname = binding->dirname;
 269   else
 270     {
 271       /* We have a relative path.  Make it absolute now.  */
 272       size_t dirname_len = strlen (binding->dirname) + 1;
 273       size_t path_max;
 274       char *ret;
 275
 276       path_max = (unsigned) PATH_MAX;
 277       path_max += 2;            /* The getcwd docs say to do this.  */
 278
 279       dirname = (char *) alloca (path_max + dirname_len);
 280       ADD_BLOCK (block_list, dirname);
 281
 282       __set_errno (0);
 283       while ((ret = getcwd (dirname, path_max)) == NULL && errno == ERANGE)
 284         {
 285           path_max += PATH_INCR;
 286           dirname = (char *) alloca (path_max + dirname_len);
 287           ADD_BLOCK (block_list, dirname);
 288           __set_errno (0);
 289         }
 290
 291       if (ret == NULL)
 292         {
 293           /* We cannot get the current working directory.  Don't signal an
 294              error but simply return the default string.  */
 295           FREE_BLOCKS (block_list);
 296           __set_errno (saved_errno);
 297           return (char *) msgid;
 298         }
 299
 300       stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
 301     }
 302
 303   /* Now determine the symbolic name of CATEGORY and its value.  */
 304   categoryname = category_to_name (category);
 305   categoryvalue = guess_category_value (category, categoryname);
 306
 307   xdomainname = (char *) alloca (strlen (categoryname)
 308                                  + strlen (domainname) + 5);
 309   ADD_BLOCK (block_list, xdomainname);
 310
 311   stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
 312                   domainname),
 313           ".mo");
 314
 315   /* Creating working area.  */
 316   single_locale = (char *) alloca (strlen (categoryvalue) + 1);
 317   ADD_BLOCK (block_list, single_locale);
 318
 319
 320   /* Search for the given string.  This is a loop because we perhaps
 321      got an ordered list of languages to consider for th translation.  */
 322   while (1)
 323     {
 324       /* Make CATEGORYVALUE point to the next element of the list.  */
 325       while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
 326         ++categoryvalue;
 327       if (categoryvalue[0] == '\0')
 328         {
 329           /* The whole contents of CATEGORYVALUE has been searched but
 330              no valid entry has been found.  We solve this situation
 331              by implicitly appending a "C" entry, i.e. no translation
 332              will take place.  */
 333           single_locale[0] = 'C';
 334           single_locale[1] = '\0';
 335         }
 336       else
 337         {
 338           char *cp = single_locale;
 339           while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
 340             *cp++ = *categoryvalue++;
 341           *cp = '\0';
 342         }
 343
 344       /* If the current locale value is C (or POSIX) we don't load a
 345          domain.  Return the MSGID.  */
 346       if (strcmp (single_locale, "C") == 0
 347           || strcmp (single_locale, "POSIX") == 0)
 348         {
 349           FREE_BLOCKS (block_list);
 350           __set_errno (saved_errno);
 351           return (char *) msgid;
 352         }
 353
 354
 355       /* Find structure describing the message catalog matching the
 356          DOMAINNAME and CATEGORY.  */
 357       domain = _nl_find_domain (dirname, single_locale, xdomainname);
 358
 359       if (domain != NULL)
 360         {
 361           retval = find_msg (domain, msgid);
 362
 363           if (retval == NULL)
 364             {
 365               int cnt;
 366
 367               for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
 368                 {
 369                   retval = find_msg (domain->successor[cnt], msgid);
 370
 371                   if (retval != NULL)
 372                     break;
 373                 }
 374             }
 375
 376           if (retval != NULL)
 377             {
 378               FREE_BLOCKS (block_list);
 379               __set_errno (saved_errno);
 380               return retval;
 381             }
 382         }
 383     }
 384   /* NOTREACHED */
 385 }
 386
 387 #ifdef _LIBC
 388 /* Alias for function name in GNU C Library.  */
 389 weak_alias (__dcgettext, dcgettext);
 390 #endif
 391
 392
 393 static char *
 394 internal_function
 395 find_msg (domain_file, msgid)
 396      struct loaded_l10nfile *domain_file;
 397      const char *msgid;
 398 {
 399   size_t top, act, bottom;
 400   struct loaded_domain *domain;
 401
 402   if (domain_file->decided == 0)
 403     _nl_load_domain (domain_file);
 404
 405   if (domain_file->data == NULL)
 406     return NULL;
 407
 408   domain = (struct loaded_domain *) domain_file->data;
 409
 410   /* Locate the MSGID and its translation.  */
 411   if (domain->hash_size > 2 && domain->hash_tab != NULL)
 412     {
 413       /* Use the hashing table.  */
 414       nls_uint32 len = strlen (msgid);
 415       nls_uint32 hash_val = hash_string (msgid);
 416       nls_uint32 idx = hash_val % domain->hash_size;
 417       nls_uint32 incr = 1 + (hash_val % (domain->hash_size - 2));
 418       nls_uint32 nstr = W (domain->must_swap, domain->hash_tab[idx]);
 419
 420       if (nstr == 0)
 421         /* Hash table entry is empty.  */
 422         return NULL;
 423
 424       if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 425           && strcmp (msgid,
 426                      domain->data + W (domain->must_swap,
 427                                        domain->orig_tab[nstr - 1].offset)) == 0)
 428         return (char *) domain->data + W (domain->must_swap,
 429                                           domain->trans_tab[nstr - 1].offset);
 430
 431       while (1)
 432         {
 433           if (idx >= domain->hash_size - incr)
 434             idx -= domain->hash_size - incr;
 435           else
 436             idx += incr;
 437
 438           nstr = W (domain->must_swap, domain->hash_tab[idx]);
 439           if (nstr == 0)
 440             /* Hash table entry is empty.  */
 441             return NULL;
 442
 443           if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 444               && strcmp (msgid,
 445                          domain->data + W (domain->must_swap,
 446                                            domain->orig_tab[nstr - 1].offset))
 447                  == 0)
 448             return (char *) domain->data
 449               + W (domain->must_swap, domain->trans_tab[nstr - 1].offset);
 450         }
 451       /* NOTREACHED */
 452     }
 453
 454   /* Now we try the default method:  binary search in the sorted
 455      array of messages.  */
 456   bottom = 0;
 457   top = domain->nstrings;
 458   while (bottom < top)
 459     {
 460       int cmp_val;
 461
 462       act = (bottom + top) / 2;
 463       cmp_val = strcmp (msgid, domain->data
 464                                + W (domain->must_swap,
 465                                     domain->orig_tab[act].offset));
 466       if (cmp_val < 0)
 467         top = act;
 468       else if (cmp_val > 0)
 469         bottom = act + 1;
 470       else
 471         break;
 472     }
 473
 474   /* If an translation is found return this.  */
 475   return bottom >= top ? NULL : (char *) domain->data
 476                                 + W (domain->must_swap,
 477                                      domain->trans_tab[act].offset);
 478 }
 479
 480
 481 /* Return string representation of locale CATEGORY.  */
 482 static const char *
 483 internal_function
 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 *
 543 internal_function
 544 guess_category_value (category, categoryname)
 545      int category;
 546      const char *categoryname;
 547 {
 548   const char *retval;
 549
 550   /* The highest priority value is the `LANGUAGE' environment
 551      variable.  This is a GNU extension.  */
 552   retval = getenv ("LANGUAGE");
 553   if (retval != NULL && retval[0] != '\0')
 554     return retval;
 555
 556   /* `LANGUAGE' is not set.  So we have to proceed with the POSIX
 557      methods of looking to `LC_ALL', `LC_xxx', and `LANG'.  On some
 558      systems this can be done by the `setlocale' function itself.  */
 559 #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
 560   return setlocale (category, NULL);
 561 #else
 562   /* Setting of LC_ALL overwrites all other.  */
 563   retval = getenv ("LC_ALL");
 564   if (retval != NULL && retval[0] != '\0')
 565     return retval;
 566
 567   /* Next comes the name of the desired category.  */
 568   retval = getenv (categoryname);
 569   if (retval != NULL && retval[0] != '\0')
 570     return retval;
 571
 572   /* Last possibility is the LANG environment variable.  */
 573   retval = getenv ("LANG");
 574   if (retval != NULL && retval[0] != '\0')
 575     return retval;
 576
 577   /* We use C as the default domain.  POSIX says this is implementation
 578      defined.  */
 579   return "C";
 580 #endif
 581 }
 582
 583 /* @@ begin of epilog @@ */
 584
 585 /* We don't want libintl.a to depend on any other library.  So we
 586    avoid the non-standard function stpcpy.  In GNU C Library this
 587    function is available, though.  Also allow the symbol HAVE_STPCPY
 588    to be defined.  */
 589 #if !_LIBC && !HAVE_STPCPY
 590 static char *
 591 stpcpy (dest, src)
 592      char *dest;
 593      const char *src;
 594 {
 595   while ((*dest++ = *src++) != '\0')
 596     /* Do nothing. */ ;
 597   return dest - 1;
 598 }
 599 #endif
 600
 601
 602 #ifdef _LIBC
 603 /* If we want to free all resources we have to do some work at
 604    program's end.  */
 605 static void __attribute__ ((unused))
 606 free_mem (void)
 607 {
 608   struct binding *runp;
 609
 610   for (runp = _nl_domain_bindings; runp != NULL; runp = runp->next)
 611     {
 612       free (runp->domainname);
 613       if (runp->dirname != _nl_default_dirname)
 614         /* Yes, this is a pointer comparison.  */
 615         free (runp->dirname);
 616     }
 617
 618   if (_nl_current_default_domain != _nl_default_default_domain)
 619     /* Yes, again a pointer comparison.  */
 620     free ((char *) _nl_current_default_domain);
 621 }
 622
 623 text_set_element (__libc_subfreeres, free_mem);
 624 #endif