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 #ifndef __set_errno
  50 # define __set_errno(val) errno = (val)
  51 #endif
  52
  53 #if defined STDC_HEADERS || defined _LIBC
  54 # include <stdlib.h>
  55 #else
  56 char *getenv ();
  57 # ifdef HAVE_MALLOC_H
  58 #  include <malloc.h>
  59 # else
  60 void free ();
  61 # endif
  62 #endif
  63
  64 #if defined HAVE_STRING_H || defined _LIBC
  65 # ifndef _GNU_SOURCE
  66 #  define _GNU_SOURCE   1
  67 # endif
  68 # include <string.h>
  69 #else
  70 # include <strings.h>
  71 #endif
  72 #if !HAVE_STRCHR && !defined _LIBC
  73 # ifndef strchr
  74 #  define strchr index
  75 # endif
  76 #endif
  77
  78 #if defined HAVE_UNISTD_H || defined _LIBC
  79 # include <unistd.h>
  80 #endif
  81
  82 #include "gettext.h"
  83 #include "gettextP.h"
  84 #ifdef _LIBC
  85 # include <libintl.h>
  86 #else
  87 # include "libgettext.h"
  88 #endif
  89 #include "hash-string.h"
  90
  91 /* @@ end of prolog @@ */
  92
  93 #ifdef _LIBC
  94 /* Rename the non ANSI C functions.  This is required by the standard
  95    because some ANSI C functions will require linking with this object
  96    file and the name space must not be polluted.  */
  97 # define getcwd __getcwd
  98 # define stpcpy __stpcpy
  99 #else
 100 # if !defined HAVE_GETCWD
 101 char *getwd ();
 102 #  define getcwd(buf, max) getwd (buf)
 103 # else
 104 char *getcwd ();
 105 # endif
 106 # ifndef HAVE_STPCPY
 107 static char *stpcpy PARAMS ((char *dest, const char *src));
 108 # endif
 109 #endif
 110
 111 /* Amount to increase buffer size by in each try.  */
 112 #define PATH_INCR 32
 113
 114 /* The following is from pathmax.h.  */
 115 /* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define
 116    PATH_MAX but might cause redefinition warnings when sys/param.h is
 117    later included (as on MORE/BSD 4.3).  */
 118 #if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && !defined(__GNUC__))
 119 # include <limits.h>
 120 #endif
 121
 122 #ifndef _POSIX_PATH_MAX
 123 # define _POSIX_PATH_MAX 255
 124 #endif
 125
 126 #if !defined(PATH_MAX) && defined(_PC_PATH_MAX)
 127 # define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX))
 128 #endif
 129
 130 /* Don't include sys/param.h if it already has been.  */
 131 #if defined(HAVE_SYS_PARAM_H) && !defined(PATH_MAX) && !defined(MAXPATHLEN)
 132 # include <sys/param.h>
 133 #endif
 134
 135 #if !defined(PATH_MAX) && defined(MAXPATHLEN)
 136 # define PATH_MAX MAXPATHLEN
 137 #endif
 138
 139 #ifndef PATH_MAX
 140 # define PATH_MAX _POSIX_PATH_MAX
 141 #endif
 142
 143 /* XPG3 defines the result of `setlocale (category, NULL)' as:
 144    ``Directs `setlocale()' to query `category' and return the current
 145      setting of `local'.''
 146    However it does not specify the exact format.  And even worse: POSIX
 147    defines this not at all.  So we can use this feature only on selected
 148    system (e.g. those using GNU C Library).  */
 149 #ifdef _LIBC
 150 # define HAVE_LOCALE_NULL
 151 #endif
 152
 153 /* Name of the default domain used for gettext(3) prior any call to
 154    textdomain(3).  The default value for this is "messages".  */
 155 const char _nl_default_default_domain[] = "messages";
 156
 157 /* Value used as the default domain for gettext(3).  */
 158 const char *_nl_current_default_domain = _nl_default_default_domain;
 159
 160 /* Contains the default location of the message catalogs.  */
 161 const char _nl_default_dirname[] = GNULOCALEDIR;
 162
 163 /* List with bindings of specific domains created by bindtextdomain()
 164    calls.  */
 165 struct binding *_nl_domain_bindings;
 166
 167 /* Prototypes for local functions.  */
 168 static char *find_msg PARAMS ((struct loaded_l10nfile *domain_file,
 169                                const char *msgid));
 170 static const char *category_to_name PARAMS ((int category));
 171 static const char *guess_category_value PARAMS ((int category,
 172                                                  const char *categoryname));
 173
 174
 175 /* For those loosing systems which don't have `alloca' we have to add
 176    some additional code emulating it.  */
 177 #ifdef HAVE_ALLOCA
 178 /* Nothing has to be done.  */
 179 # define ADD_BLOCK(list, address) /* nothing */
 180 # define FREE_BLOCKS(list) /* nothing */
 181 #else
 182 struct block_list
 183 {
 184   void *address;
 185   struct block_list *next;
 186 };
 187 # define ADD_BLOCK(list, addr)                                                \
 188   do {                                                                        \
 189     struct block_list *newp = (struct block_list *) malloc (sizeof (*newp));  \
 190     /* If we cannot get a free block we cannot add the new element to         \
 191        the list.  */                                                          \
 192     if (newp != NULL) {                                                       \
 193       newp->address = (addr);                                                 \
 194       newp->next = (list);                                                    \
 195       (list) = newp;                                                          \
 196     }                                                                         \
 197   } while (0)
 198 # define FREE_BLOCKS(list)                                                    \
 199   do {                                                                        \
 200     while (list != NULL) {                                                    \
 201       struct block_list *old = list;                                          \
 202       list = list->next;                                                      \
 203       free (old);                                                             \
 204     }                                                                         \
 205   } while (0)
 206 # undef alloca
 207 # define alloca(size) (malloc (size))
 208 #endif  /* have alloca */
 209
 210
 211 /* Names for the libintl functions are a problem.  They must not clash
 212    with existing names and they should follow ANSI C.  But this source
 213    code is also used in GNU C Library where the names have a __
 214    prefix.  So we have to make a difference here.  */
 215 #ifdef _LIBC
 216 # define DCGETTEXT __dcgettext
 217 #else
 218 # define DCGETTEXT dcgettext__
 219 #endif
 220
 221 /* Look up MSGID in the DOMAINNAME message catalog for the current CATEGORY
 222    locale.  */
 223 char *
 224 DCGETTEXT (domainname, msgid, category)
 225      const char *domainname;
 226      const char *msgid;
 227      int category;
 228 {
 229 #ifndef HAVE_ALLOCA
 230   struct block_list *block_list = NULL;
 231 #endif
 232   struct loaded_l10nfile *domain;
 233   struct binding *binding;
 234   const char *categoryname;
 235   const char *categoryvalue;
 236   char *dirname, *xdomainname;
 237   char *single_locale;
 238   char *retval;
 239   int saved_errno = errno;
 240
 241   /* If no real MSGID is given return NULL.  */
 242   if (msgid == NULL)
 243     return NULL;
 244
 245   /* If DOMAINNAME is NULL, we are interested in the default domain.  If
 246      CATEGORY is not LC_MESSAGES this might not make much sense but the
 247      defintion left this undefined.  */
 248   if (domainname == NULL)
 249     domainname = _nl_current_default_domain;
 250
 251   /* First find matching binding.  */
 252   for (binding = _nl_domain_bindings; binding != NULL; binding = binding->next)
 253     {
 254       int compare = strcmp (domainname, binding->domainname);
 255       if (compare == 0)
 256         /* We found it!  */
 257         break;
 258       if (compare < 0)
 259         {
 260           /* It is not in the list.  */
 261           binding = NULL;
 262           break;
 263         }
 264     }
 265
 266   if (binding == NULL)
 267     dirname = (char *) _nl_default_dirname;
 268   else if (binding->dirname[0] == '/')
 269     dirname = binding->dirname;
 270   else
 271     {
 272       /* We have a relative path.  Make it absolute now.  */
 273       size_t dirname_len = strlen (binding->dirname) + 1;
 274       size_t path_max;
 275       char *ret;
 276
 277       path_max = (unsigned) PATH_MAX;
 278       path_max += 2;            /* The getcwd docs say to do this.  */
 279
 280       dirname = (char *) alloca (path_max + dirname_len);
 281       ADD_BLOCK (block_list, dirname);
 282
 283       __set_errno (0);
 284       while ((ret = getcwd (dirname, path_max)) == NULL && errno == ERANGE)
 285         {
 286           path_max += PATH_INCR;
 287           dirname = (char *) alloca (path_max + dirname_len);
 288           ADD_BLOCK (block_list, dirname);
 289           __set_errno (0);
 290         }
 291
 292       if (ret == NULL)
 293         {
 294           /* We cannot get the current working directory.  Don't signal an
 295              error but simply return the default string.  */
 296           FREE_BLOCKS (block_list);
 297           __set_errno (saved_errno);
 298           return (char *) msgid;
 299         }
 300
 301       /* We don't want libintl.a to depend on any other library.  So
 302          we avoid the non-standard function stpcpy.  In GNU C Library
 303          this function is available, though.  Also allow the symbol
 304          HAVE_STPCPY to be defined.  */
 305       stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
 306     }
 307
 308   /* Now determine the symbolic name of CATEGORY and its value.  */
 309   categoryname = category_to_name (category);
 310   categoryvalue = guess_category_value (category, categoryname);
 311
 312   xdomainname = (char *) alloca (strlen (categoryname)
 313                                  + strlen (domainname) + 5);
 314   ADD_BLOCK (block_list, xdomainname);
 315   /* We don't want libintl.a to depend on any other library.  So we
 316      avoid the non-standard function stpcpy.  In GNU C Library this
 317      function is available, though.  Also allow the symbol HAVE_STPCPY
 318      to be defined.  */
 319   stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
 320                   domainname),
 321           ".mo");
 322
 323   /* Creating working area.  */
 324   single_locale = (char *) alloca (strlen (categoryvalue) + 1);
 325   ADD_BLOCK (block_list, single_locale);
 326
 327
 328   /* Search for the given string.  This is a loop because we perhaps
 329      got an ordered list of languages to consider for th translation.  */
 330   while (1)
 331     {
 332       /* Make CATEGORYVALUE point to the next element of the list.  */
 333       while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
 334         ++categoryvalue;
 335       if (categoryvalue[0] == '\0')
 336         {
 337           /* The whole contents of CATEGORYVALUE has been searched but
 338              no valid entry has been found.  We solve this situation
 339              by implicitely appending a "C" entry, i.e. no translation
 340              will take place.  */
 341           single_locale[0] = 'C';
 342           single_locale[1] = '\0';
 343         }
 344       else
 345         {
 346           char *cp = single_locale;
 347           while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
 348             *cp++ = *categoryvalue++;
 349           *cp = '\0';
 350         }
 351
 352       /* If the current locale value is C (or POSIX) we don't load a
 353          domain.  Return the MSGID.  */
 354       if (strcmp (single_locale, "C") == 0
 355           || strcmp (single_locale, "POSIX") == 0)
 356         {
 357           FREE_BLOCKS (block_list);
 358           __set_errno (saved_errno);
 359           return (char *) msgid;
 360         }
 361
 362
 363       /* Find structure describing the message catalog matching the
 364          DOMAINNAME and CATEGORY.  */
 365       domain = _nl_find_domain (dirname, single_locale, xdomainname);
 366
 367       if (domain != NULL)
 368         {
 369           retval = find_msg (domain, msgid);
 370
 371           if (retval == NULL)
 372             {
 373               int cnt;
 374
 375               for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
 376                 {
 377                   retval = find_msg (domain->successor[cnt], msgid);
 378
 379                   if (retval != NULL)
 380                     break;
 381                 }
 382             }
 383
 384           if (retval != NULL)
 385             {
 386               FREE_BLOCKS (block_list);
 387               __set_errno (saved_errno);
 388               return retval;
 389             }
 390         }
 391     }
 392   /* NOTREACHED */
 393 }
 394
 395 #ifdef _LIBC
 396 /* Alias for function name in GNU C Library.  */
 397 weak_alias (__dcgettext, dcgettext);
 398 #endif
 399
 400
 401 static char *
 402 find_msg (domain_file, msgid)
 403      struct loaded_l10nfile *domain_file;
 404      const char *msgid;
 405 {
 406   size_t top, act, bottom;
 407   struct loaded_domain *domain;
 408
 409   if (domain_file->decided == 0)
 410     _nl_load_domain (domain_file);
 411
 412   if (domain_file->data == NULL)
 413     return NULL;
 414
 415   domain = (struct loaded_domain *) domain_file->data;
 416
 417   /* Locate the MSGID and its translation.  */
 418   if (domain->hash_size > 2 && domain->hash_tab != NULL)
 419     {
 420       /* Use the hashing table.  */
 421       nls_uint32 len = strlen (msgid);
 422       nls_uint32 hash_val = hash_string (msgid);
 423       nls_uint32 idx = hash_val % domain->hash_size;
 424       nls_uint32 incr = 1 + (hash_val % (domain->hash_size - 2));
 425       nls_uint32 nstr = W (domain->must_swap, domain->hash_tab[idx]);
 426
 427       if (nstr == 0)
 428         /* Hash table entry is empty.  */
 429         return NULL;
 430
 431       if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 432           && strcmp (msgid,
 433                      domain->data + W (domain->must_swap,
 434                                        domain->orig_tab[nstr - 1].offset)) == 0)
 435         return (char *) domain->data + W (domain->must_swap,
 436                                           domain->trans_tab[nstr - 1].offset);
 437
 438       while (1)
 439         {
 440           if (idx >= domain->hash_size - incr)
 441             idx -= domain->hash_size - incr;
 442           else
 443             idx += incr;
 444
 445           nstr = W (domain->must_swap, domain->hash_tab[idx]);
 446           if (nstr == 0)
 447             /* Hash table entry is empty.  */
 448             return NULL;
 449
 450           if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 451               && strcmp (msgid,
 452                          domain->data + W (domain->must_swap,
 453                                            domain->orig_tab[nstr - 1].offset))
 454                  == 0)
 455             return (char *) domain->data
 456               + W (domain->must_swap, domain->trans_tab[nstr - 1].offset);
 457         }
 458       /* NOTREACHED */
 459     }
 460
 461   /* Now we try the default method:  binary search in the sorted
 462      array of messages.  */
 463   bottom = 0;
 464   top = domain->nstrings;
 465   while (bottom < top)
 466     {
 467       int cmp_val;
 468
 469       act = (bottom + top) / 2;
 470       cmp_val = strcmp (msgid, domain->data
 471                                + W (domain->must_swap,
 472                                     domain->orig_tab[act].offset));
 473       if (cmp_val < 0)
 474         top = act;
 475       else if (cmp_val > 0)
 476         bottom = act + 1;
 477       else
 478         break;
 479     }
 480
 481   /* If an translation is found return this.  */
 482   return bottom >= top ? NULL : (char *) domain->data
 483                                 + W (domain->must_swap,
 484                                      domain->trans_tab[act].offset);
 485 }
 486
 487
 488 /* Return string representation of locale CATEGORY.  */
 489 static const char *
 490 category_to_name (category)
 491      int category;
 492 {
 493   const char *retval;
 494
 495   switch (category)
 496   {
 497 #ifdef LC_COLLATE
 498   case LC_COLLATE:
 499     retval = "LC_COLLATE";
 500     break;
 501 #endif
 502 #ifdef LC_CTYPE
 503   case LC_CTYPE:
 504     retval = "LC_CTYPE";
 505     break;
 506 #endif
 507 #ifdef LC_MONETARY
 508   case LC_MONETARY:
 509     retval = "LC_MONETARY";
 510     break;
 511 #endif
 512 #ifdef LC_NUMERIC
 513   case LC_NUMERIC:
 514     retval = "LC_NUMERIC";
 515     break;
 516 #endif
 517 #ifdef LC_TIME
 518   case LC_TIME:
 519     retval = "LC_TIME";
 520     break;
 521 #endif
 522 #ifdef LC_MESSAGES
 523   case LC_MESSAGES:
 524     retval = "LC_MESSAGES";
 525     break;
 526 #endif
 527 #ifdef LC_RESPONSE
 528   case LC_RESPONSE:
 529     retval = "LC_RESPONSE";
 530     break;
 531 #endif
 532 #ifdef LC_ALL
 533   case LC_ALL:
 534     /* This might not make sense but is perhaps better than any other
 535        value.  */
 536     retval = "LC_ALL";
 537     break;
 538 #endif
 539   default:
 540     /* If you have a better idea for a default value let me know.  */
 541     retval = "LC_XXX";
 542   }
 543
 544   return retval;
 545 }
 546
 547 /* Guess value of current locale from value of the environment variables.  */
 548 static const char *guess_category_value (category, categoryname)
 549      int category;
 550      const char *categoryname;
 551 {
 552   const char *retval;
 553
 554   /* The highest priority value is the `LANGUAGE' environment
 555      variable.  This is a GNU extension.  */
 556   retval = getenv ("LANGUAGE");
 557   if (retval != NULL && retval[0] != '\0')
 558     return retval;
 559
 560   /* `LANGUAGE' is not set.  So we have to proceed with the POSIX
 561      methods of looking to `LC_ALL', `LC_xxx', and `LANG'.  On some
 562      systems this can be done by the `setlocale' function itself.  */
 563 #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
 564   return setlocale (category, NULL);
 565 #else
 566   /* Setting of LC_ALL overwrites all other.  */
 567   retval = getenv ("LC_ALL");
 568   if (retval != NULL && retval[0] != '\0')
 569     return retval;
 570
 571   /* Next comes the name of the desired category.  */
 572   retval = getenv (categoryname);
 573   if (retval != NULL && retval[0] != '\0')
 574     return retval;
 575
 576   /* Last possibility is the LANG environment variable.  */
 577   retval = getenv ("LANG");
 578   if (retval != NULL && retval[0] != '\0')
 579     return retval;
 580
 581   /* We use C as the default domain.  POSIX says this is implementation
 582      defined.  */
 583   return "C";
 584 #endif
 585 }
 586
 587 /* @@ begin of epilog @@ */
 588
 589 /* We don't want libintl.a to depend on any other library.  So we
 590    avoid the non-standard function stpcpy.  In GNU C Library this
 591    function is available, though.  Also allow the symbol HAVE_STPCPY
 592    to be defined.  */
 593 #if !_LIBC && !HAVE_STPCPY
 594 static char *
 595 stpcpy (dest, src)
 596      char *dest;
 597      const char *src;
 598 {
 599   while ((*dest++ = *src++) != '\0')
 600     /* Do nothing. */ ;
 601   return dest - 1;
 602 }
 603 #endif