intl/dcgettext.c

   1 /* Implementation of the dcgettext(3) function.
   2    Copyright (C) 1995, 1996, 1997, 1998, 1999 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 #if defined __GNUC__ && !defined C_ALLOCA
  29 # define alloca __builtin_alloca
  30 # define HAVE_ALLOCA 1
  31 #else
  32 # if (defined HAVE_ALLOCA_H || defined _LIBC) && !defined C_ALLOCA
  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 #if defined HAVE_LOCALE_H || defined _LIBC
  83 # include <locale.h>
  84 #endif
  85
  86 #include "gettext.h"
  87 #include "gettextP.h"
  88 #ifdef _LIBC
  89 # include <libintl.h>
  90 #else
  91 # include "libgettext.h"
  92 #endif
  93 #include "hash-string.h"
  94
  95 /* @@ end of prolog @@ */
  96
  97 #ifdef _LIBC
  98 /* Rename the non ANSI C functions.  This is required by the standard
  99    because some ANSI C functions will require linking with this object
 100    file and the name space must not be polluted.  */
 101 # define getcwd __getcwd
 102 # ifndef stpcpy
 103 #  define stpcpy __stpcpy
 104 # endif
 105 #else
 106 # if !defined HAVE_GETCWD
 107 char *getwd ();
 108 #  define getcwd(buf, max) getwd (buf)
 109 # else
 110 char *getcwd ();
 111 # endif
 112 # ifndef HAVE_STPCPY
 113 static char *stpcpy PARAMS ((char *dest, const char *src));
 114 # endif
 115 #endif
 116
 117 /* Amount to increase buffer size by in each try.  */
 118 #define PATH_INCR 32
 119
 120 /* The following is from pathmax.h.  */
 121 /* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define
 122    PATH_MAX but might cause redefinition warnings when sys/param.h is
 123    later included (as on MORE/BSD 4.3).  */
 124 #if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && !defined(__GNUC__))
 125 # include <limits.h>
 126 #endif
 127
 128 #ifndef _POSIX_PATH_MAX
 129 # define _POSIX_PATH_MAX 255
 130 #endif
 131
 132 #if !defined(PATH_MAX) && defined(_PC_PATH_MAX)
 133 # define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX))
 134 #endif
 135
 136 /* Don't include sys/param.h if it already has been.  */
 137 #if defined(HAVE_SYS_PARAM_H) && !defined(PATH_MAX) && !defined(MAXPATHLEN)
 138 # include <sys/param.h>
 139 #endif
 140
 141 #if !defined(PATH_MAX) && defined(MAXPATHLEN)
 142 # define PATH_MAX MAXPATHLEN
 143 #endif
 144
 145 #ifndef PATH_MAX
 146 # define PATH_MAX _POSIX_PATH_MAX
 147 #endif
 148
 149 /* XPG3 defines the result of `setlocale (category, NULL)' as:
 150    ``Directs `setlocale()' to query `category' and return the current
 151      setting of `local'.''
 152    However it does not specify the exact format.  And even worse: POSIX
 153    defines this not at all.  So we can use this feature only on selected
 154    system (e.g. those using GNU C Library).  */
 155 #ifdef _LIBC
 156 # define HAVE_LOCALE_NULL
 157 #endif
 158
 159 /* Name of the default domain used for gettext(3) prior any call to
 160    textdomain(3).  The default value for this is "messages".  */
 161 const char _nl_default_default_domain[] = "messages";
 162
 163 /* Value used as the default domain for gettext(3).  */
 164 const char *_nl_current_default_domain = _nl_default_default_domain;
 165
 166 /* Contains the default location of the message catalogs.  */
 167 const char _nl_default_dirname[] = GNULOCALEDIR;
 168
 169 /* List with bindings of specific domains created by bindtextdomain()
 170    calls.  */
 171 struct binding *_nl_domain_bindings;
 172
 173 /* Prototypes for local functions.  */
 174 static char *find_msg PARAMS ((struct loaded_l10nfile *domain_file,
 175                                const char *msgid)) internal_function;
 176 static const char *category_to_name PARAMS ((int category)) internal_function;
 177 static const char *guess_category_value PARAMS ((int category,
 178                                                  const char *categoryname))
 179      internal_function;
 180
 181
 182 /* For those loosing systems which don't have `alloca' we have to add
 183    some additional code emulating it.  */
 184 #ifdef HAVE_ALLOCA
 185 /* Nothing has to be done.  */
 186 # define ADD_BLOCK(list, address) /* nothing */
 187 # define FREE_BLOCKS(list) /* nothing */
 188 #else
 189 struct block_list
 190 {
 191   void *address;
 192   struct block_list *next;
 193 };
 194 # define ADD_BLOCK(list, addr)                                                \
 195   do {                                                                        \
 196     struct block_list *newp = (struct block_list *) malloc (sizeof (*newp));  \
 197     /* If we cannot get a free block we cannot add the new element to         \
 198        the list.  */                                                          \
 199     if (newp != NULL) {                                                       \
 200       newp->address = (addr);                                                 \
 201       newp->next = (list);                                                    \
 202       (list) = newp;                                                          \
 203     }                                                                         \
 204   } while (0)
 205 # define FREE_BLOCKS(list)                                                    \
 206   do {                                                                        \
 207     while (list != NULL) {                                                    \
 208       struct block_list *old = list;                                          \
 209       list = list->next;                                                      \
 210       free (old);                                                             \
 211     }                                                                         \
 212   } while (0)
 213 # undef alloca
 214 # define alloca(size) (malloc (size))
 215 #endif  /* have alloca */
 216
 217
 218 /* Names for the libintl functions are a problem.  They must not clash
 219    with existing names and they should follow ANSI C.  But this source
 220    code is also used in GNU C Library where the names have a __
 221    prefix.  So we have to make a difference here.  */
 222 #ifdef _LIBC
 223 # define DCGETTEXT __dcgettext
 224 #else
 225 # define DCGETTEXT dcgettext__
 226 #endif
 227
 228 /* Checking whether the binaries runs SUID must be done and glibc provides
 229    easier methods therefore we make a difference here.  */
 230 #ifdef _LIBC
 231 # define ENABLE_SECURE __libc_enable_secure
 232 # define DETERMINE_SECURE
 233 #else
 234 static int enable_secure;
 235 # define ENABLE_SECURE (enable_secure == 1)
 236 # define DETERMINE_SECURE \
 237   if (enable_secure == 0)                                                     \
 238     {                                                                         \
 239       if (getuid () != geteuid () || getgid () != getegid ())                 \
 240         enable_secure = 1;                                                    \
 241       else                                                                    \
 242         enable_secure = -1;                                                   \
 243     }
 244 #endif
 245
 246 /* Look up MSGID in the DOMAINNAME message catalog for the current CATEGORY
 247    locale.  */
 248 char *
 249 DCGETTEXT (domainname, msgid, category)
 250      const char *domainname;
 251      const char *msgid;
 252      int category;
 253 {
 254 #ifndef HAVE_ALLOCA
 255   struct block_list *block_list = NULL;
 256 #endif
 257   struct loaded_l10nfile *domain;
 258   struct binding *binding;
 259   const char *categoryname;
 260   const char *categoryvalue;
 261   char *dirname, *xdomainname;
 262   char *single_locale;
 263   char *retval;
 264   int saved_errno = errno;
 265
 266   /* If no real MSGID is given return NULL.  */
 267   if (msgid == NULL)
 268     return NULL;
 269
 270   /* See whether this is a SUID binary or not.  */
 271   DETERMINE_SECURE;
 272
 273   /* If DOMAINNAME is NULL, we are interested in the default domain.  If
 274      CATEGORY is not LC_MESSAGES this might not make much sense but the
 275      definition left this undefined.  */
 276   if (domainname == NULL)
 277     domainname = _nl_current_default_domain;
 278
 279   /* First find matching binding.  */
 280   for (binding = _nl_domain_bindings; binding != NULL; binding = binding->next)
 281     {
 282       int compare = strcmp (domainname, binding->domainname);
 283       if (compare == 0)
 284         /* We found it!  */
 285         break;
 286       if (compare < 0)
 287         {
 288           /* It is not in the list.  */
 289           binding = NULL;
 290           break;
 291         }
 292     }
 293
 294   if (binding == NULL)
 295     dirname = (char *) _nl_default_dirname;
 296   else if (binding->dirname[0] == '/')
 297     dirname = binding->dirname;
 298   else
 299     {
 300       /* We have a relative path.  Make it absolute now.  */
 301       size_t dirname_len = strlen (binding->dirname) + 1;
 302       size_t path_max;
 303       char *ret;
 304
 305       path_max = (unsigned int) PATH_MAX;
 306       path_max += 2;            /* The getcwd docs say to do this.  */
 307
 308       dirname = (char *) alloca (path_max + dirname_len);
 309       ADD_BLOCK (block_list, dirname);
 310
 311       __set_errno (0);
 312       while ((ret = getcwd (dirname, path_max)) == NULL && errno == ERANGE)
 313         {
 314           path_max += PATH_INCR;
 315           dirname = (char *) alloca (path_max + dirname_len);
 316           ADD_BLOCK (block_list, dirname);
 317           __set_errno (0);
 318         }
 319
 320       if (ret == NULL)
 321         {
 322           /* We cannot get the current working directory.  Don't signal an
 323              error but simply return the default string.  */
 324           FREE_BLOCKS (block_list);
 325           __set_errno (saved_errno);
 326           return (char *) msgid;
 327         }
 328
 329       stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
 330     }
 331
 332   /* Now determine the symbolic name of CATEGORY and its value.  */
 333   categoryname = category_to_name (category);
 334   categoryvalue = guess_category_value (category, categoryname);
 335
 336   xdomainname = (char *) alloca (strlen (categoryname)
 337                                  + strlen (domainname) + 5);
 338   ADD_BLOCK (block_list, xdomainname);
 339
 340   stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
 341                   domainname),
 342           ".mo");
 343
 344   /* Creating working area.  */
 345   single_locale = (char *) alloca (strlen (categoryvalue) + 1);
 346   ADD_BLOCK (block_list, single_locale);
 347
 348
 349   /* Search for the given string.  This is a loop because we perhaps
 350      got an ordered list of languages to consider for the translation.  */
 351   while (1)
 352     {
 353       /* Make CATEGORYVALUE point to the next element of the list.  */
 354       while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
 355         ++categoryvalue;
 356       if (categoryvalue[0] == '\0')
 357         {
 358           /* The whole contents of CATEGORYVALUE has been searched but
 359              no valid entry has been found.  We solve this situation
 360              by implicitly appending a "C" entry, i.e. no translation
 361              will take place.  */
 362           single_locale[0] = 'C';
 363           single_locale[1] = '\0';
 364         }
 365       else
 366         {
 367           char *cp = single_locale;
 368           while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
 369             *cp++ = *categoryvalue++;
 370           *cp = '\0';
 371
 372           /* When this is a SUID binary we must not allow accessing files
 373              outside the dedicated directories.  */
 374           if (ENABLE_SECURE
 375               && (memchr (single_locale, '/',
 376                           _nl_find_language (single_locale) - single_locale)
 377                   != NULL))
 378             /* Ingore this entry.  */
 379             continue;
 380         }
 381
 382       /* If the current locale value is C (or POSIX) we don't load a
 383          domain.  Return the MSGID.  */
 384       if (strcmp (single_locale, "C") == 0
 385           || strcmp (single_locale, "POSIX") == 0)
 386         {
 387           FREE_BLOCKS (block_list);
 388           __set_errno (saved_errno);
 389           return (char *) msgid;
 390         }
 391
 392
 393       /* Find structure describing the message catalog matching the
 394          DOMAINNAME and CATEGORY.  */
 395       domain = _nl_find_domain (dirname, single_locale, xdomainname);
 396
 397       if (domain != NULL)
 398         {
 399           retval = find_msg (domain, msgid);
 400
 401           if (retval == NULL)
 402             {
 403               int cnt;
 404
 405               for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
 406                 {
 407                   retval = find_msg (domain->successor[cnt], msgid);
 408
 409                   if (retval != NULL)
 410                     break;
 411                 }
 412             }
 413
 414           if (retval != NULL)
 415             {
 416               FREE_BLOCKS (block_list);
 417               __set_errno (saved_errno);
 418               return retval;
 419             }
 420         }
 421     }
 422   /* NOTREACHED */
 423 }
 424
 425 #ifdef _LIBC
 426 /* Alias for function name in GNU C Library.  */
 427 weak_alias (__dcgettext, dcgettext);
 428 #endif
 429
 430
 431 static char *
 432 internal_function
 433 find_msg (domain_file, msgid)
 434      struct loaded_l10nfile *domain_file;
 435      const char *msgid;
 436 {
 437   size_t act = 0;
 438   size_t top, bottom;
 439   struct loaded_domain *domain;
 440
 441   if (domain_file->decided == 0)
 442     _nl_load_domain (domain_file);
 443
 444   if (domain_file->data == NULL)
 445     return NULL;
 446
 447   domain = (struct loaded_domain *) domain_file->data;
 448
 449   /* Locate the MSGID and its translation.  */
 450   if (domain->hash_size > 2 && domain->hash_tab != NULL)
 451     {
 452       /* Use the hashing table.  */
 453       nls_uint32 len = strlen (msgid);
 454       nls_uint32 hash_val = hash_string (msgid);
 455       nls_uint32 idx = hash_val % domain->hash_size;
 456       nls_uint32 incr = 1 + (hash_val % (domain->hash_size - 2));
 457       nls_uint32 nstr = W (domain->must_swap, domain->hash_tab[idx]);
 458
 459       if (nstr == 0)
 460         /* Hash table entry is empty.  */
 461         return NULL;
 462
 463       if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 464           && strcmp (msgid,
 465                      domain->data + W (domain->must_swap,
 466                                        domain->orig_tab[nstr - 1].offset)) == 0)
 467         return (char *) domain->data + W (domain->must_swap,
 468                                           domain->trans_tab[nstr - 1].offset);
 469
 470       while (1)
 471         {
 472           if (idx >= domain->hash_size - incr)
 473             idx -= domain->hash_size - incr;
 474           else
 475             idx += incr;
 476
 477           nstr = W (domain->must_swap, domain->hash_tab[idx]);
 478           if (nstr == 0)
 479             /* Hash table entry is empty.  */
 480             return NULL;
 481
 482           if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 483               && strcmp (msgid,
 484                          domain->data + W (domain->must_swap,
 485                                            domain->orig_tab[nstr - 1].offset))
 486                  == 0)
 487             return (char *) domain->data
 488               + W (domain->must_swap, domain->trans_tab[nstr - 1].offset);
 489         }
 490       /* NOTREACHED */
 491     }
 492
 493   /* Now we try the default method:  binary search in the sorted
 494      array of messages.  */
 495   bottom = 0;
 496   top = domain->nstrings;
 497   while (bottom < top)
 498     {
 499       int cmp_val;
 500
 501       act = (bottom + top) / 2;
 502       cmp_val = strcmp (msgid, domain->data
 503                                + W (domain->must_swap,
 504                                     domain->orig_tab[act].offset));
 505       if (cmp_val < 0)
 506         top = act;
 507       else if (cmp_val > 0)
 508         bottom = act + 1;
 509       else
 510         break;
 511     }
 512
 513   /* If an translation is found return this.  */
 514   return bottom >= top ? NULL : (char *) domain->data
 515                                 + W (domain->must_swap,
 516                                      domain->trans_tab[act].offset);
 517 }
 518
 519
 520 /* Return string representation of locale CATEGORY.  */
 521 static const char *
 522 internal_function
 523 category_to_name (category)
 524      int category;
 525 {
 526   const char *retval;
 527
 528   switch (category)
 529   {
 530 #ifdef LC_COLLATE
 531   case LC_COLLATE:
 532     retval = "LC_COLLATE";
 533     break;
 534 #endif
 535 #ifdef LC_CTYPE
 536   case LC_CTYPE:
 537     retval = "LC_CTYPE";
 538     break;
 539 #endif
 540 #ifdef LC_MONETARY
 541   case LC_MONETARY:
 542     retval = "LC_MONETARY";
 543     break;
 544 #endif
 545 #ifdef LC_NUMERIC
 546   case LC_NUMERIC:
 547     retval = "LC_NUMERIC";
 548     break;
 549 #endif
 550 #ifdef LC_TIME
 551   case LC_TIME:
 552     retval = "LC_TIME";
 553     break;
 554 #endif
 555 #ifdef LC_MESSAGES
 556   case LC_MESSAGES:
 557     retval = "LC_MESSAGES";
 558     break;
 559 #endif
 560 #ifdef LC_RESPONSE
 561   case LC_RESPONSE:
 562     retval = "LC_RESPONSE";
 563     break;
 564 #endif
 565 #ifdef LC_ALL
 566   case LC_ALL:
 567     /* This might not make sense but is perhaps better than any other
 568        value.  */
 569     retval = "LC_ALL";
 570     break;
 571 #endif
 572   default:
 573     /* If you have a better idea for a default value let me know.  */
 574     retval = "LC_XXX";
 575   }
 576
 577   return retval;
 578 }
 579
 580 /* Guess value of current locale from value of the environment variables.  */
 581 static const char *
 582 internal_function
 583 guess_category_value (category, categoryname)
 584      int category;
 585      const char *categoryname;
 586 {
 587   const char *retval;
 588
 589   /* The highest priority value is the `LANGUAGE' environment
 590      variable.  This is a GNU extension.  */
 591   retval = getenv ("LANGUAGE");
 592   if (retval != NULL && retval[0] != '\0')
 593     return retval;
 594
 595   /* `LANGUAGE' is not set.  So we have to proceed with the POSIX
 596      methods of looking to `LC_ALL', `LC_xxx', and `LANG'.  On some
 597      systems this can be done by the `setlocale' function itself.  */
 598 #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
 599   return setlocale (category, NULL);
 600 #else
 601   /* Setting of LC_ALL overwrites all other.  */
 602   retval = getenv ("LC_ALL");
 603   if (retval != NULL && retval[0] != '\0')
 604     return retval;
 605
 606   /* Next comes the name of the desired category.  */
 607   retval = getenv (categoryname);
 608   if (retval != NULL && retval[0] != '\0')
 609     return retval;
 610
 611   /* Last possibility is the LANG environment variable.  */
 612   retval = getenv ("LANG");
 613   if (retval != NULL && retval[0] != '\0')
 614     return retval;
 615
 616   /* We use C as the default domain.  POSIX says this is implementation
 617      defined.  */
 618   return "C";
 619 #endif
 620 }
 621
 622 /* @@ begin of epilog @@ */
 623
 624 /* We don't want libintl.a to depend on any other library.  So we
 625    avoid the non-standard function stpcpy.  In GNU C Library this
 626    function is available, though.  Also allow the symbol HAVE_STPCPY
 627    to be defined.  */
 628 #if !_LIBC && !HAVE_STPCPY
 629 static char *
 630 stpcpy (dest, src)
 631      char *dest;
 632      const char *src;
 633 {
 634   while ((*dest++ = *src++) != '\0')
 635     /* Do nothing. */ ;
 636   return dest - 1;
 637 }
 638 #endif
 639
 640
 641 #ifdef _LIBC
 642 /* If we want to free all resources we have to do some work at
 643    program's end.  */
 644 static void __attribute__ ((unused))
 645 free_mem (void)
 646 {
 647   struct binding *runp;
 648
 649   for (runp = _nl_domain_bindings; runp != NULL; runp = runp->next)
 650     {
 651       free (runp->domainname);
 652       if (runp->dirname != _nl_default_dirname)
 653         /* Yes, this is a pointer comparison.  */
 654         free (runp->dirname);
 655     }
 656
 657   if (_nl_current_default_domain != _nl_default_default_domain)
 658     /* Yes, again a pointer comparison.  */
 659     free ((char *) _nl_current_default_domain);
 660 }
 661
 662 text_set_element (__libc_subfreeres, free_mem);
 663 #endif