gcc/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 # ifndef _GNU_SOURCE
  23 #  define _GNU_SOURCE   1
  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 # 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 /* The internal variables in the standalone libintl.a must have different
  89    names than the internal variables in GNU libc, otherwise programs
  90    using libintl.a cannot be linked statically.  */
  91 #if !defined _LIBC
  92 # define _nl_default_default_domain _nl_default_default_domain__
  93 # define _nl_current_default_domain _nl_current_default_domain__
  94 # define _nl_default_dirname _nl_default_dirname__
  95 # define _nl_domain_bindings _nl_domain_bindings__
  96 #endif
  97
  98 /* @@ end of prolog @@ */
  99
 100 #ifdef _LIBC
 101 /* Rename the non ANSI C functions.  This is required by the standard
 102    because some ANSI C functions will require linking with this object
 103    file and the name space must not be polluted.  */
 104 # define getcwd __getcwd
 105 # ifndef stpcpy
 106 #  define stpcpy __stpcpy
 107 # endif
 108 #else
 109 # if !defined HAVE_GETCWD
 110 char *getwd ();
 111 #  define getcwd(buf, max) getwd (buf)
 112 # elif !defined (HAVE_DECL_GETCWD)
 113 char *getcwd ();
 114 # endif
 115 # ifndef HAVE_STPCPY
 116 static char *stpcpy PARAMS ((char *dest, const char *src));
 117 # endif
 118 #endif
 119
 120 /* Amount to increase buffer size by in each try.  */
 121 #define PATH_INCR 32
 122
 123 /* The following is from pathmax.h.  */
 124 /* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define
 125    PATH_MAX but might cause redefinition warnings when sys/param.h is
 126    later included (as on MORE/BSD 4.3).  */
 127 #if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && !defined(__GNUC__))
 128 # include <limits.h>
 129 #endif
 130
 131 #ifndef _POSIX_PATH_MAX
 132 # define _POSIX_PATH_MAX 255
 133 #endif
 134
 135 #if !defined(PATH_MAX) && defined(_PC_PATH_MAX)
 136 # define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX))
 137 #endif
 138
 139 /* Don't include sys/param.h if it already has been.  */
 140 #if defined(HAVE_SYS_PARAM_H) && !defined(PATH_MAX) && !defined(MAXPATHLEN)
 141 # include <sys/param.h>
 142 #endif
 143
 144 #if !defined(PATH_MAX) && defined(MAXPATHLEN)
 145 # define PATH_MAX MAXPATHLEN
 146 #endif
 147
 148 #ifndef PATH_MAX
 149 # define PATH_MAX _POSIX_PATH_MAX
 150 #endif
 151
 152 /* XPG3 defines the result of `setlocale (category, NULL)' as:
 153    ``Directs `setlocale()' to query `category' and return the current
 154      setting of `local'.''
 155    However it does not specify the exact format.  And even worse: POSIX
 156    defines this not at all.  So we can use this feature only on selected
 157    system (e.g. those using GNU C Library).  */
 158 #ifdef _LIBC
 159 # define HAVE_LOCALE_NULL
 160 #endif
 161
 162 /* Name of the default domain used for gettext(3) prior any call to
 163    textdomain(3).  The default value for this is "messages".  */
 164 const char _nl_default_default_domain[] = "messages";
 165
 166 /* Value used as the default domain for gettext(3).  */
 167 const char *_nl_current_default_domain = _nl_default_default_domain;
 168
 169 /* Contains the default location of the message catalogs.  */
 170 const char _nl_default_dirname[] = GNULOCALEDIR;
 171
 172 /* List with bindings of specific domains created by bindtextdomain()
 173    calls.  */
 174 struct binding *_nl_domain_bindings;
 175
 176 /* Prototypes for local functions.  */
 177 static char *find_msg PARAMS ((struct loaded_l10nfile *domain_file,
 178                                const char *msgid)) internal_function;
 179 static const char *category_to_name PARAMS ((int category)) internal_function;
 180 static const char *guess_category_value PARAMS ((int category,
 181                                                  const char *categoryname))
 182      internal_function;
 183
 184
 185 /* For those loosing systems which don't have `alloca' we have to add
 186    some additional code emulating it.  */
 187 #ifdef HAVE_ALLOCA
 188 /* Nothing has to be done.  */
 189 # define ADD_BLOCK(list, address) /* nothing */
 190 # define FREE_BLOCKS(list) /* nothing */
 191 #else
 192 struct block_list
 193 {
 194   void *address;
 195   struct block_list *next;
 196 };
 197 # define ADD_BLOCK(list, addr)                                                \
 198   do {                                                                        \
 199     struct block_list *newp = (struct block_list *) malloc (sizeof (*newp));  \
 200     /* If we cannot get a free block we cannot add the new element to         \
 201        the list.  */                                                          \
 202     if (newp != NULL) {                                                       \
 203       newp->address = (addr);                                                 \
 204       newp->next = (list);                                                    \
 205       (list) = newp;                                                          \
 206     }                                                                         \
 207   } while (0)
 208 # define FREE_BLOCKS(list)                                                    \
 209   do {                                                                        \
 210     while (list != NULL) {                                                    \
 211       struct block_list *old = list;                                          \
 212       list = list->next;                                                      \
 213       free (old);                                                             \
 214     }                                                                         \
 215   } while (0)
 216 # undef alloca
 217 # define alloca(size) (malloc (size))
 218 #endif  /* have alloca */
 219
 220
 221 /* Names for the libintl functions are a problem.  They must not clash
 222    with existing names and they should follow ANSI C.  But this source
 223    code is also used in GNU C Library where the names have a __
 224    prefix.  So we have to make a difference here.  */
 225 #ifdef _LIBC
 226 # define DCGETTEXT __dcgettext
 227 #else
 228 # define DCGETTEXT dcgettext__
 229 #endif
 230
 231 /* Look up MSGID in the DOMAINNAME message catalog for the current CATEGORY
 232    locale.  */
 233 char *
 234 DCGETTEXT (domainname, msgid, category)
 235      const char *domainname;
 236      const char *msgid;
 237      int category;
 238 {
 239 #ifndef HAVE_ALLOCA
 240   struct block_list *block_list = NULL;
 241 #endif
 242   struct loaded_l10nfile *domain;
 243   struct binding *binding;
 244   const char *categoryname;
 245   const char *categoryvalue;
 246   char *dirname, *xdomainname;
 247   char *single_locale;
 248   char *retval;
 249   int saved_errno = errno;
 250
 251   /* If no real MSGID is given return NULL.  */
 252   if (msgid == NULL)
 253     return NULL;
 254
 255   /* If DOMAINNAME is NULL, we are interested in the default domain.  If
 256      CATEGORY is not LC_MESSAGES this might not make much sense but the
 257      defintion left this undefined.  */
 258   if (domainname == NULL)
 259     domainname = _nl_current_default_domain;
 260
 261   /* First find matching binding.  */
 262   for (binding = _nl_domain_bindings; binding != NULL; binding = binding->next)
 263     {
 264       int compare = strcmp (domainname, binding->domainname);
 265       if (compare == 0)
 266         /* We found it!  */
 267         break;
 268       if (compare < 0)
 269         {
 270           /* It is not in the list.  */
 271           binding = NULL;
 272           break;
 273         }
 274     }
 275
 276   if (binding == NULL)
 277     dirname = (char *) _nl_default_dirname;
 278   else if (binding->dirname[0] == '/')
 279     dirname = binding->dirname;
 280   else
 281     {
 282       /* We have a relative path.  Make it absolute now.  */
 283       size_t dirname_len = strlen (binding->dirname) + 1;
 284       size_t path_max;
 285       char *ret;
 286
 287       path_max = (unsigned) PATH_MAX;
 288       path_max += 2;            /* The getcwd docs say to do this.  */
 289
 290       dirname = (char *) alloca (path_max + dirname_len);
 291       ADD_BLOCK (block_list, dirname);
 292
 293       __set_errno (0);
 294       while ((ret = getcwd (dirname, path_max)) == NULL && errno == ERANGE)
 295         {
 296           path_max += PATH_INCR;
 297           dirname = (char *) alloca (path_max + dirname_len);
 298           ADD_BLOCK (block_list, dirname);
 299           __set_errno (0);
 300         }
 301
 302       if (ret == NULL)
 303         {
 304           /* We cannot get the current working directory.  Don't signal an
 305              error but simply return the default string.  */
 306           FREE_BLOCKS (block_list);
 307           __set_errno (saved_errno);
 308           return (char *) msgid;
 309         }
 310
 311       stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
 312     }
 313
 314   /* Now determine the symbolic name of CATEGORY and its value.  */
 315   categoryname = category_to_name (category);
 316   categoryvalue = guess_category_value (category, categoryname);
 317
 318   xdomainname = (char *) alloca (strlen (categoryname)
 319                                  + strlen (domainname) + 5);
 320   ADD_BLOCK (block_list, xdomainname);
 321
 322   stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
 323                   domainname),
 324           ".mo");
 325
 326   /* Creating working area.  */
 327   single_locale = (char *) alloca (strlen (categoryvalue) + 1);
 328   ADD_BLOCK (block_list, single_locale);
 329
 330
 331   /* Search for the given string.  This is a loop because we perhaps
 332      got an ordered list of languages to consider for th translation.  */
 333   while (1)
 334     {
 335       /* Make CATEGORYVALUE point to the next element of the list.  */
 336       while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
 337         ++categoryvalue;
 338       if (categoryvalue[0] == '\0')
 339         {
 340           /* The whole contents of CATEGORYVALUE has been searched but
 341              no valid entry has been found.  We solve this situation
 342              by implicitly appending a "C" entry, i.e. no translation
 343              will take place.  */
 344           single_locale[0] = 'C';
 345           single_locale[1] = '\0';
 346         }
 347       else
 348         {
 349           char *cp = single_locale;
 350           while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
 351             *cp++ = *categoryvalue++;
 352           *cp = '\0';
 353         }
 354
 355       /* If the current locale value is C (or POSIX) we don't load a
 356          domain.  Return the MSGID.  */
 357       if (strcmp (single_locale, "C") == 0
 358           || strcmp (single_locale, "POSIX") == 0)
 359         {
 360           FREE_BLOCKS (block_list);
 361           __set_errno (saved_errno);
 362           return (char *) msgid;
 363         }
 364
 365
 366       /* Find structure describing the message catalog matching the
 367          DOMAINNAME and CATEGORY.  */
 368       domain = _nl_find_domain (dirname, single_locale, xdomainname);
 369
 370       if (domain != NULL)
 371         {
 372           retval = find_msg (domain, msgid);
 373
 374           if (retval == NULL)
 375             {
 376               int cnt;
 377
 378               for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
 379                 {
 380                   retval = find_msg (domain->successor[cnt], msgid);
 381
 382                   if (retval != NULL)
 383                     break;
 384                 }
 385             }
 386
 387           if (retval != NULL)
 388             {
 389               FREE_BLOCKS (block_list);
 390               __set_errno (saved_errno);
 391               return retval;
 392             }
 393         }
 394     }
 395   /* NOTREACHED */
 396 }
 397
 398 #ifdef _LIBC
 399 /* Alias for function name in GNU C Library.  */
 400 weak_alias (__dcgettext, dcgettext);
 401 #endif
 402
 403
 404 static char *
 405 internal_function
 406 find_msg (domain_file, msgid)
 407      struct loaded_l10nfile *domain_file;
 408      const char *msgid;
 409 {
 410   size_t top, act, bottom;
 411   struct loaded_domain *domain;
 412
 413   if (domain_file->decided == 0)
 414     _nl_load_domain (domain_file);
 415
 416   if (domain_file->data == NULL)
 417     return NULL;
 418
 419   domain = (struct loaded_domain *) domain_file->data;
 420
 421   /* Locate the MSGID and its translation.  */
 422   if (domain->hash_size > 2 && domain->hash_tab != NULL)
 423     {
 424       /* Use the hashing table.  */
 425       nls_uint32 len = strlen (msgid);
 426       nls_uint32 hash_val = hash_string (msgid);
 427       nls_uint32 idx = hash_val % domain->hash_size;
 428       nls_uint32 incr = 1 + (hash_val % (domain->hash_size - 2));
 429       nls_uint32 nstr = W (domain->must_swap, domain->hash_tab[idx]);
 430
 431       if (nstr == 0)
 432         /* Hash table entry is empty.  */
 433         return NULL;
 434
 435       if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 436           && strcmp (msgid,
 437                      domain->data + W (domain->must_swap,
 438                                        domain->orig_tab[nstr - 1].offset)) == 0)
 439         return (char *) domain->data + W (domain->must_swap,
 440                                           domain->trans_tab[nstr - 1].offset);
 441
 442       while (1)
 443         {
 444           if (idx >= domain->hash_size - incr)
 445             idx -= domain->hash_size - incr;
 446           else
 447             idx += incr;
 448
 449           nstr = W (domain->must_swap, domain->hash_tab[idx]);
 450           if (nstr == 0)
 451             /* Hash table entry is empty.  */
 452             return NULL;
 453
 454           if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 455               && strcmp (msgid,
 456                          domain->data + W (domain->must_swap,
 457                                            domain->orig_tab[nstr - 1].offset))
 458                  == 0)
 459             return (char *) domain->data
 460               + W (domain->must_swap, domain->trans_tab[nstr - 1].offset);
 461         }
 462       /* NOTREACHED */
 463     }
 464
 465   /* Now we try the default method:  binary search in the sorted
 466      array of messages.  */
 467   bottom = 0;
 468   act = 0;
 469   top = domain->nstrings;
 470   while (bottom < top)
 471     {
 472       int cmp_val;
 473
 474       act = (bottom + top) / 2;
 475       cmp_val = strcmp (msgid, domain->data
 476                                + W (domain->must_swap,
 477                                     domain->orig_tab[act].offset));
 478       if (cmp_val < 0)
 479         top = act;
 480       else if (cmp_val > 0)
 481         bottom = act + 1;
 482       else
 483         break;
 484     }
 485
 486   /* If an translation is found return this.  */
 487   return bottom >= top ? NULL : (char *) domain->data
 488                                 + W (domain->must_swap,
 489                                      domain->trans_tab[act].offset);
 490 }
 491
 492
 493 /* Return string representation of locale CATEGORY.  */
 494 static const char *
 495 internal_function
 496 category_to_name (category)
 497      int category;
 498 {
 499   const char *retval;
 500
 501   switch (category)
 502   {
 503 #ifdef LC_COLLATE
 504   case LC_COLLATE:
 505     retval = "LC_COLLATE";
 506     break;
 507 #endif
 508 #ifdef LC_CTYPE
 509   case LC_CTYPE:
 510     retval = "LC_CTYPE";
 511     break;
 512 #endif
 513 #ifdef LC_MONETARY
 514   case LC_MONETARY:
 515     retval = "LC_MONETARY";
 516     break;
 517 #endif
 518 #ifdef LC_NUMERIC
 519   case LC_NUMERIC:
 520     retval = "LC_NUMERIC";
 521     break;
 522 #endif
 523 #ifdef LC_TIME
 524   case LC_TIME:
 525     retval = "LC_TIME";
 526     break;
 527 #endif
 528 #ifdef LC_MESSAGES
 529   case LC_MESSAGES:
 530     retval = "LC_MESSAGES";
 531     break;
 532 #endif
 533 #ifdef LC_RESPONSE
 534   case LC_RESPONSE:
 535     retval = "LC_RESPONSE";
 536     break;
 537 #endif
 538 #ifdef LC_ALL
 539   case LC_ALL:
 540     /* This might not make sense but is perhaps better than any other
 541        value.  */
 542     retval = "LC_ALL";
 543     break;
 544 #endif
 545   default:
 546     /* If you have a better idea for a default value let me know.  */
 547     retval = "LC_XXX";
 548   }
 549
 550   return retval;
 551 }
 552
 553 /* Guess value of current locale from value of the environment variables.  */
 554 static const char *
 555 internal_function
 556 guess_category_value (category, categoryname)
 557      int category;
 558      const char *categoryname;
 559 {
 560   const char *retval;
 561   (void) category;  /* shut up compiler */
 562   (void) categoryname;  /* ditto */
 563
 564   /* The highest priority value is the `LANGUAGE' environment
 565      variable.  This is a GNU extension.  */
 566   retval = getenv ("LANGUAGE");
 567   if (retval != NULL && retval[0] != '\0')
 568     return retval;
 569
 570   /* `LANGUAGE' is not set.  So we have to proceed with the POSIX
 571      methods of looking to `LC_ALL', `LC_xxx', and `LANG'.  On some
 572      systems this can be done by the `setlocale' function itself.  */
 573 #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
 574   return setlocale (category, NULL);
 575 #else
 576   /* Setting of LC_ALL overwrites all other.  */
 577   retval = getenv ("LC_ALL");
 578   if (retval != NULL && retval[0] != '\0')
 579     return retval;
 580
 581   /* Next comes the name of the desired category.  */
 582   retval = getenv (categoryname);
 583   if (retval != NULL && retval[0] != '\0')
 584     return retval;
 585
 586   /* Last possibility is the LANG environment variable.  */
 587   retval = getenv ("LANG");
 588   if (retval != NULL && retval[0] != '\0')
 589     return retval;
 590
 591   /* We use C as the default domain.  POSIX says this is implementation
 592      defined.  */
 593   return "C";
 594 #endif
 595 }
 596
 597 /* @@ begin of epilog @@ */
 598
 599 /* We don't want libintl.a to depend on any other library.  So we
 600    avoid the non-standard function stpcpy.  In GNU C Library this
 601    function is available, though.  Also allow the symbol HAVE_STPCPY
 602    to be defined.  */
 603 #if !_LIBC && !HAVE_STPCPY
 604 static char *
 605 stpcpy (dest, src)
 606      char *dest;
 607      const char *src;
 608 {
 609   while ((*dest++ = *src++) != '\0')
 610     /* Do nothing. */ ;
 611   return dest - 1;
 612 }
 613 #endif
 614
 615
 616 #ifdef _LIBC
 617 /* If we want to free all resources we have to do some work at
 618    program's end.  */
 619 static void __attribute__ ((unused))
 620 free_mem (void)
 621 {
 622   struct binding *runp;
 623
 624   for (runp = _nl_domain_bindings; runp != NULL; runp = runp->next)
 625     {
 626       free (runp->domainname);
 627       if (runp->dirname != _nl_default_dirname)
 628         /* Yes, this is a pointer comparison.  */
 629         free (runp->dirname);
 630     }
 631
 632   if (_nl_current_default_domain != _nl_default_default_domain)
 633     /* Yes, again a pointer comparison.  */
 634     free ((char *) _nl_current_default_domain);
 635 }
 636
 637 text_set_element (__libc_subfreeres, free_mem);
 638 #endif