intl/dcgettext.c

   1 /* Implementation of the dcgettext(3) function.
   2    Copyright (C) 1995, 1996, 1997, 1998 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.
   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 not,
  19    write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  20    Boston, MA 02111-1307, 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 # ifndef stpcpy
  99 #  define stpcpy __stpcpy
 100 # endif
 101 #else
 102 # if !defined HAVE_GETCWD
 103 char *getwd ();
 104 #  define getcwd(buf, max) getwd (buf)
 105 # else
 106 char *getcwd ();
 107 # endif
 108 # ifndef HAVE_STPCPY
 109 static char *stpcpy PARAMS ((char *dest, const char *src));
 110 # endif
 111 #endif
 112
 113 /* Amount to increase buffer size by in each try.  */
 114 #define PATH_INCR 32
 115
 116 /* The following is from pathmax.h.  */
 117 /* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define
 118    PATH_MAX but might cause redefinition warnings when sys/param.h is
 119    later included (as on MORE/BSD 4.3).  */
 120 #if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && !defined(__GNUC__))
 121 # include <limits.h>
 122 #endif
 123
 124 #ifndef _POSIX_PATH_MAX
 125 # define _POSIX_PATH_MAX 255
 126 #endif
 127
 128 #if !defined(PATH_MAX) && defined(_PC_PATH_MAX)
 129 # define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX))
 130 #endif
 131
 132 /* Don't include sys/param.h if it already has been.  */
 133 #if defined(HAVE_SYS_PARAM_H) && !defined(PATH_MAX) && !defined(MAXPATHLEN)
 134 # include <sys/param.h>
 135 #endif
 136
 137 #if !defined(PATH_MAX) && defined(MAXPATHLEN)
 138 # define PATH_MAX MAXPATHLEN
 139 #endif
 140
 141 #ifndef PATH_MAX
 142 # define PATH_MAX _POSIX_PATH_MAX
 143 #endif
 144
 145 /* XPG3 defines the result of `setlocale (category, NULL)' as:
 146    ``Directs `setlocale()' to query `category' and return the current
 147      setting of `local'.''
 148    However it does not specify the exact format.  And even worse: POSIX
 149    defines this not at all.  So we can use this feature only on selected
 150    system (e.g. those using GNU C Library).  */
 151 #ifdef _LIBC
 152 # define HAVE_LOCALE_NULL
 153 #endif
 154
 155 /* Name of the default domain used for gettext(3) prior any call to
 156    textdomain(3).  The default value for this is "messages".  */
 157 const char _nl_default_default_domain[] = "messages";
 158
 159 /* Value used as the default domain for gettext(3).  */
 160 const char *_nl_current_default_domain = _nl_default_default_domain;
 161
 162 /* Contains the default location of the message catalogs.  */
 163 const char _nl_default_dirname[] = GNULOCALEDIR;
 164
 165 /* List with bindings of specific domains created by bindtextdomain()
 166    calls.  */
 167 struct binding *_nl_domain_bindings;
 168
 169 /* Prototypes for local functions.  */
 170 static char *find_msg PARAMS ((struct loaded_l10nfile *domain_file,
 171                                const char *msgid)) internal_function;
 172 static const char *category_to_name PARAMS ((int category)) internal_function;
 173 static const char *guess_category_value PARAMS ((int category,
 174                                                  const char *categoryname))
 175      internal_function;
 176
 177
 178 /* For those loosing systems which don't have `alloca' we have to add
 179    some additional code emulating it.  */
 180 #ifdef HAVE_ALLOCA
 181 /* Nothing has to be done.  */
 182 # define ADD_BLOCK(list, address) /* nothing */
 183 # define FREE_BLOCKS(list) /* nothing */
 184 #else
 185 struct block_list
 186 {
 187   void *address;
 188   struct block_list *next;
 189 };
 190 # define ADD_BLOCK(list, addr)                                                \
 191   do {                                                                        \
 192     struct block_list *newp = (struct block_list *) malloc (sizeof (*newp));  \
 193     /* If we cannot get a free block we cannot add the new element to         \
 194        the list.  */                                                          \
 195     if (newp != NULL) {                                                       \
 196       newp->address = (addr);                                                 \
 197       newp->next = (list);                                                    \
 198       (list) = newp;                                                          \
 199     }                                                                         \
 200   } while (0)
 201 # define FREE_BLOCKS(list)                                                    \
 202   do {                                                                        \
 203     while (list != NULL) {                                                    \
 204       struct block_list *old = list;                                          \
 205       list = list->next;                                                      \
 206       free (old);                                                             \
 207     }                                                                         \
 208   } while (0)
 209 # undef alloca
 210 # define alloca(size) (malloc (size))
 211 #endif  /* have alloca */
 212
 213
 214 /* Names for the libintl functions are a problem.  They must not clash
 215    with existing names and they should follow ANSI C.  But this source
 216    code is also used in GNU C Library where the names have a __
 217    prefix.  So we have to make a difference here.  */
 218 #ifdef _LIBC
 219 # define DCGETTEXT __dcgettext
 220 #else
 221 # define DCGETTEXT dcgettext__
 222 #endif
 223
 224 /* Look up MSGID in the DOMAINNAME message catalog for the current CATEGORY
 225    locale.  */
 226 char *
 227 DCGETTEXT (domainname, msgid, category)
 228      const char *domainname;
 229      const char *msgid;
 230      int category;
 231 {
 232 #ifndef HAVE_ALLOCA
 233   struct block_list *block_list = NULL;
 234 #endif
 235   struct loaded_l10nfile *domain;
 236   struct binding *binding;
 237   const char *categoryname;
 238   const char *categoryvalue;
 239   char *dirname, *xdomainname;
 240   char *single_locale;
 241   char *retval;
 242   int saved_errno = errno;
 243
 244   /* If no real MSGID is given return NULL.  */
 245   if (msgid == NULL)
 246     return NULL;
 247
 248   /* If DOMAINNAME is NULL, we are interested in the default domain.  If
 249      CATEGORY is not LC_MESSAGES this might not make much sense but the
 250      defintion left this undefined.  */
 251   if (domainname == NULL)
 252     domainname = _nl_current_default_domain;
 253
 254   /* First find matching binding.  */
 255   for (binding = _nl_domain_bindings; binding != NULL; binding = binding->next)
 256     {
 257       int compare = strcmp (domainname, binding->domainname);
 258       if (compare == 0)
 259         /* We found it!  */
 260         break;
 261       if (compare < 0)
 262         {
 263           /* It is not in the list.  */
 264           binding = NULL;
 265           break;
 266         }
 267     }
 268
 269   if (binding == NULL)
 270     dirname = (char *) _nl_default_dirname;
 271   else if (binding->dirname[0] == '/')
 272     dirname = binding->dirname;
 273   else
 274     {
 275       /* We have a relative path.  Make it absolute now.  */
 276       size_t dirname_len = strlen (binding->dirname) + 1;
 277       size_t path_max;
 278       char *ret;
 279
 280       path_max = (unsigned) PATH_MAX;
 281       path_max += 2;            /* The getcwd docs say to do this.  */
 282
 283       dirname = (char *) alloca (path_max + dirname_len);
 284       ADD_BLOCK (block_list, dirname);
 285
 286       __set_errno (0);
 287       while ((ret = getcwd (dirname, path_max)) == NULL && errno == ERANGE)
 288         {
 289           path_max += PATH_INCR;
 290           dirname = (char *) alloca (path_max + dirname_len);
 291           ADD_BLOCK (block_list, dirname);
 292           __set_errno (0);
 293         }
 294
 295       if (ret == NULL)
 296         {
 297           /* We cannot get the current working directory.  Don't signal an
 298              error but simply return the default string.  */
 299           FREE_BLOCKS (block_list);
 300           __set_errno (saved_errno);
 301           return (char *) msgid;
 302         }
 303
 304       stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
 305     }
 306
 307   /* Now determine the symbolic name of CATEGORY and its value.  */
 308   categoryname = category_to_name (category);
 309   categoryvalue = guess_category_value (category, categoryname);
 310
 311   xdomainname = (char *) alloca (strlen (categoryname)
 312                                  + strlen (domainname) + 5);
 313   ADD_BLOCK (block_list, xdomainname);
 314
 315   stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
 316                   domainname),
 317           ".mo");
 318
 319   /* Creating working area.  */
 320   single_locale = (char *) alloca (strlen (categoryvalue) + 1);
 321   ADD_BLOCK (block_list, single_locale);
 322
 323
 324   /* Search for the given string.  This is a loop because we perhaps
 325      got an ordered list of languages to consider for th translation.  */
 326   while (1)
 327     {
 328       /* Make CATEGORYVALUE point to the next element of the list.  */
 329       while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
 330         ++categoryvalue;
 331       if (categoryvalue[0] == '\0')
 332         {
 333           /* The whole contents of CATEGORYVALUE has been searched but
 334              no valid entry has been found.  We solve this situation
 335              by implicitly appending a "C" entry, i.e. no translation
 336              will take place.  */
 337           single_locale[0] = 'C';
 338           single_locale[1] = '\0';
 339         }
 340       else
 341         {
 342           char *cp = single_locale;
 343           while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
 344             *cp++ = *categoryvalue++;
 345           *cp = '\0';
 346         }
 347
 348       /* If the current locale value is C (or POSIX) we don't load a
 349          domain.  Return the MSGID.  */
 350       if (strcmp (single_locale, "C") == 0
 351           || strcmp (single_locale, "POSIX") == 0)
 352         {
 353           FREE_BLOCKS (block_list);
 354           __set_errno (saved_errno);
 355           return (char *) msgid;
 356         }
 357
 358
 359       /* Find structure describing the message catalog matching the
 360          DOMAINNAME and CATEGORY.  */
 361       domain = _nl_find_domain (dirname, single_locale, xdomainname);
 362
 363       if (domain != NULL)
 364         {
 365           retval = find_msg (domain, msgid);
 366
 367           if (retval == NULL)
 368             {
 369               int cnt;
 370
 371               for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
 372                 {
 373                   retval = find_msg (domain->successor[cnt], msgid);
 374
 375                   if (retval != NULL)
 376                     break;
 377                 }
 378             }
 379
 380           if (retval != NULL)
 381             {
 382               FREE_BLOCKS (block_list);
 383               __set_errno (saved_errno);
 384               return retval;
 385             }
 386         }
 387     }
 388   /* NOTREACHED */
 389 }
 390
 391 #ifdef _LIBC
 392 /* Alias for function name in GNU C Library.  */
 393 weak_alias (__dcgettext, dcgettext);
 394 #endif
 395
 396
 397 static char *
 398 internal_function
 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 internal_function
 488 category_to_name (category)
 489      int category;
 490 {
 491   const char *retval;
 492
 493   switch (category)
 494   {
 495 #ifdef LC_COLLATE
 496   case LC_COLLATE:
 497     retval = "LC_COLLATE";
 498     break;
 499 #endif
 500 #ifdef LC_CTYPE
 501   case LC_CTYPE:
 502     retval = "LC_CTYPE";
 503     break;
 504 #endif
 505 #ifdef LC_MONETARY
 506   case LC_MONETARY:
 507     retval = "LC_MONETARY";
 508     break;
 509 #endif
 510 #ifdef LC_NUMERIC
 511   case LC_NUMERIC:
 512     retval = "LC_NUMERIC";
 513     break;
 514 #endif
 515 #ifdef LC_TIME
 516   case LC_TIME:
 517     retval = "LC_TIME";
 518     break;
 519 #endif
 520 #ifdef LC_MESSAGES
 521   case LC_MESSAGES:
 522     retval = "LC_MESSAGES";
 523     break;
 524 #endif
 525 #ifdef LC_RESPONSE
 526   case LC_RESPONSE:
 527     retval = "LC_RESPONSE";
 528     break;
 529 #endif
 530 #ifdef LC_ALL
 531   case LC_ALL:
 532     /* This might not make sense but is perhaps better than any other
 533        value.  */
 534     retval = "LC_ALL";
 535     break;
 536 #endif
 537   default:
 538     /* If you have a better idea for a default value let me know.  */
 539     retval = "LC_XXX";
 540   }
 541
 542   return retval;
 543 }
 544
 545 /* Guess value of current locale from value of the environment variables.  */
 546 static const char *
 547 internal_function
 548 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
 604
 605
 606 #ifdef _LIBC
 607 /* If we want to free all resources we have to do some work at
 608    program's end.  */
 609 static void __attribute__ ((unused))
 610 free_mem (void)
 611 {
 612   struct binding *runp;
 613
 614   for (runp = _nl_domain_bindings; runp != NULL; runp = runp->next)
 615     {
 616       free (runp->domainname);
 617       if (runp->dirname != _nl_default_dirname)
 618         /* Yes, this is a pointer comparison.  */
 619         free (runp->dirname);
 620     }
 621
 622   if (_nl_current_default_domain != _nl_default_default_domain)
 623     /* Yes, again a pointer comparison.  */
 624     free ((char *) _nl_current_default_domain);
 625 }
 626
 627 text_set_element (__libc_subfreeres, free_mem);
 628 #endif