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 #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 #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 /* Checking whether the binaries runs SUID must be done and glibc provides
 225    easier methods therefore we make a difference here.  */
 226 #ifdef _LIBC
 227 # define ENABLE_SECURE __libc_enable_secure
 228 # define DETERMINE_SECURE
 229 #else
 230 static int enable_secure;
 231 # define ENABLE_SECURE (enable_secure == 1)
 232 # define DETERMINE_SECURE \
 233   if (enable_secure == 0)                                                     \
 234     {                                                                         \
 235       if (getuid () != geteuid () || getgid () != getegid ())                 \
 236         enable_secure = 1;                                                    \
 237       else                                                                    \
 238         enable_secure = -1;                                                   \
 239     }
 240 #endif
 241
 242 /* Look up MSGID in the DOMAINNAME message catalog for the current CATEGORY
 243    locale.  */
 244 char *
 245 DCGETTEXT (domainname, msgid, category)
 246      const char *domainname;
 247      const char *msgid;
 248      int category;
 249 {
 250 #ifndef HAVE_ALLOCA
 251   struct block_list *block_list = NULL;
 252 #endif
 253   struct loaded_l10nfile *domain;
 254   struct binding *binding;
 255   const char *categoryname;
 256   const char *categoryvalue;
 257   char *dirname, *xdomainname;
 258   char *single_locale;
 259   char *retval;
 260   int saved_errno = errno;
 261
 262   /* If no real MSGID is given return NULL.  */
 263   if (msgid == NULL)
 264     return NULL;
 265
 266   /* See whether this is a SUID binary or not.  */
 267   DETERMINE_SECURE;
 268
 269   /* If DOMAINNAME is NULL, we are interested in the default domain.  If
 270      CATEGORY is not LC_MESSAGES this might not make much sense but the
 271      definition left this undefined.  */
 272   if (domainname == NULL)
 273     domainname = _nl_current_default_domain;
 274
 275   /* First find matching binding.  */
 276   for (binding = _nl_domain_bindings; binding != NULL; binding = binding->next)
 277     {
 278       int compare = strcmp (domainname, binding->domainname);
 279       if (compare == 0)
 280         /* We found it!  */
 281         break;
 282       if (compare < 0)
 283         {
 284           /* It is not in the list.  */
 285           binding = NULL;
 286           break;
 287         }
 288     }
 289
 290   if (binding == NULL)
 291     dirname = (char *) _nl_default_dirname;
 292   else if (binding->dirname[0] == '/')
 293     dirname = binding->dirname;
 294   else
 295     {
 296       /* We have a relative path.  Make it absolute now.  */
 297       size_t dirname_len = strlen (binding->dirname) + 1;
 298       size_t path_max;
 299       char *ret;
 300
 301       path_max = (unsigned int) PATH_MAX;
 302       path_max += 2;            /* The getcwd docs say to do this.  */
 303
 304       dirname = (char *) alloca (path_max + dirname_len);
 305       ADD_BLOCK (block_list, dirname);
 306
 307       __set_errno (0);
 308       while ((ret = getcwd (dirname, path_max)) == NULL && errno == ERANGE)
 309         {
 310           path_max += PATH_INCR;
 311           dirname = (char *) alloca (path_max + dirname_len);
 312           ADD_BLOCK (block_list, dirname);
 313           __set_errno (0);
 314         }
 315
 316       if (ret == NULL)
 317         {
 318           /* We cannot get the current working directory.  Don't signal an
 319              error but simply return the default string.  */
 320           FREE_BLOCKS (block_list);
 321           __set_errno (saved_errno);
 322           return (char *) msgid;
 323         }
 324
 325       stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
 326     }
 327
 328   /* Now determine the symbolic name of CATEGORY and its value.  */
 329   categoryname = category_to_name (category);
 330   categoryvalue = guess_category_value (category, categoryname);
 331
 332   xdomainname = (char *) alloca (strlen (categoryname)
 333                                  + strlen (domainname) + 5);
 334   ADD_BLOCK (block_list, xdomainname);
 335
 336   stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
 337                   domainname),
 338           ".mo");
 339
 340   /* Creating working area.  */
 341   single_locale = (char *) alloca (strlen (categoryvalue) + 1);
 342   ADD_BLOCK (block_list, single_locale);
 343
 344
 345   /* Search for the given string.  This is a loop because we perhaps
 346      got an ordered list of languages to consider for the translation.  */
 347   while (1)
 348     {
 349       /* Make CATEGORYVALUE point to the next element of the list.  */
 350       while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
 351         ++categoryvalue;
 352       if (categoryvalue[0] == '\0')
 353         {
 354           /* The whole contents of CATEGORYVALUE has been searched but
 355              no valid entry has been found.  We solve this situation
 356              by implicitly appending a "C" entry, i.e. no translation
 357              will take place.  */
 358           single_locale[0] = 'C';
 359           single_locale[1] = '\0';
 360         }
 361       else
 362         {
 363           char *cp = single_locale;
 364           while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
 365             *cp++ = *categoryvalue++;
 366           *cp = '\0';
 367
 368           /* When this is a SUID binary we must not allow accessing files
 369              outside the dedicated directories.  */
 370           if (ENABLE_SECURE
 371               && (memchr (single_locale, '/',
 372                           _nl_find_language (single_locale) - single_locale)
 373                   != NULL))
 374             /* Ingore this entry.  */
 375             continue;
 376         }
 377
 378       /* If the current locale value is C (or POSIX) we don't load a
 379          domain.  Return the MSGID.  */
 380       if (strcmp (single_locale, "C") == 0
 381           || strcmp (single_locale, "POSIX") == 0)
 382         {
 383           FREE_BLOCKS (block_list);
 384           __set_errno (saved_errno);
 385           return (char *) msgid;
 386         }
 387
 388
 389       /* Find structure describing the message catalog matching the
 390          DOMAINNAME and CATEGORY.  */
 391       domain = _nl_find_domain (dirname, single_locale, xdomainname);
 392
 393       if (domain != NULL)
 394         {
 395           retval = find_msg (domain, msgid);
 396
 397           if (retval == NULL)
 398             {
 399               int cnt;
 400
 401               for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
 402                 {
 403                   retval = find_msg (domain->successor[cnt], msgid);
 404
 405                   if (retval != NULL)
 406                     break;
 407                 }
 408             }
 409
 410           if (retval != NULL)
 411             {
 412               FREE_BLOCKS (block_list);
 413               __set_errno (saved_errno);
 414               return retval;
 415             }
 416         }
 417     }
 418   /* NOTREACHED */
 419 }
 420
 421 #ifdef _LIBC
 422 /* Alias for function name in GNU C Library.  */
 423 weak_alias (__dcgettext, dcgettext);
 424 #endif
 425
 426
 427 static char *
 428 internal_function
 429 find_msg (domain_file, msgid)
 430      struct loaded_l10nfile *domain_file;
 431      const char *msgid;
 432 {
 433   size_t act = 0;
 434   size_t top, bottom;
 435   struct loaded_domain *domain;
 436
 437   if (domain_file->decided == 0)
 438     _nl_load_domain (domain_file);
 439
 440   if (domain_file->data == NULL)
 441     return NULL;
 442
 443   domain = (struct loaded_domain *) domain_file->data;
 444
 445   /* Locate the MSGID and its translation.  */
 446   if (domain->hash_size > 2 && domain->hash_tab != NULL)
 447     {
 448       /* Use the hashing table.  */
 449       nls_uint32 len = strlen (msgid);
 450       nls_uint32 hash_val = hash_string (msgid);
 451       nls_uint32 idx = hash_val % domain->hash_size;
 452       nls_uint32 incr = 1 + (hash_val % (domain->hash_size - 2));
 453       nls_uint32 nstr = W (domain->must_swap, domain->hash_tab[idx]);
 454
 455       if (nstr == 0)
 456         /* Hash table entry is empty.  */
 457         return NULL;
 458
 459       if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 460           && strcmp (msgid,
 461                      domain->data + W (domain->must_swap,
 462                                        domain->orig_tab[nstr - 1].offset)) == 0)
 463         return (char *) domain->data + W (domain->must_swap,
 464                                           domain->trans_tab[nstr - 1].offset);
 465
 466       while (1)
 467         {
 468           if (idx >= domain->hash_size - incr)
 469             idx -= domain->hash_size - incr;
 470           else
 471             idx += incr;
 472
 473           nstr = W (domain->must_swap, domain->hash_tab[idx]);
 474           if (nstr == 0)
 475             /* Hash table entry is empty.  */
 476             return NULL;
 477
 478           if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 479               && strcmp (msgid,
 480                          domain->data + W (domain->must_swap,
 481                                            domain->orig_tab[nstr - 1].offset))
 482                  == 0)
 483             return (char *) domain->data
 484               + W (domain->must_swap, domain->trans_tab[nstr - 1].offset);
 485         }
 486       /* NOTREACHED */
 487     }
 488
 489   /* Now we try the default method:  binary search in the sorted
 490      array of messages.  */
 491   bottom = 0;
 492   top = domain->nstrings;
 493   while (bottom < top)
 494     {
 495       int cmp_val;
 496
 497       act = (bottom + top) / 2;
 498       cmp_val = strcmp (msgid, domain->data
 499                                + W (domain->must_swap,
 500                                     domain->orig_tab[act].offset));
 501       if (cmp_val < 0)
 502         top = act;
 503       else if (cmp_val > 0)
 504         bottom = act + 1;
 505       else
 506         break;
 507     }
 508
 509   /* If an translation is found return this.  */
 510   return bottom >= top ? NULL : (char *) domain->data
 511                                 + W (domain->must_swap,
 512                                      domain->trans_tab[act].offset);
 513 }
 514
 515
 516 /* Return string representation of locale CATEGORY.  */
 517 static const char *
 518 internal_function
 519 category_to_name (category)
 520      int category;
 521 {
 522   const char *retval;
 523
 524   switch (category)
 525   {
 526 #ifdef LC_COLLATE
 527   case LC_COLLATE:
 528     retval = "LC_COLLATE";
 529     break;
 530 #endif
 531 #ifdef LC_CTYPE
 532   case LC_CTYPE:
 533     retval = "LC_CTYPE";
 534     break;
 535 #endif
 536 #ifdef LC_MONETARY
 537   case LC_MONETARY:
 538     retval = "LC_MONETARY";
 539     break;
 540 #endif
 541 #ifdef LC_NUMERIC
 542   case LC_NUMERIC:
 543     retval = "LC_NUMERIC";
 544     break;
 545 #endif
 546 #ifdef LC_TIME
 547   case LC_TIME:
 548     retval = "LC_TIME";
 549     break;
 550 #endif
 551 #ifdef LC_MESSAGES
 552   case LC_MESSAGES:
 553     retval = "LC_MESSAGES";
 554     break;
 555 #endif
 556 #ifdef LC_RESPONSE
 557   case LC_RESPONSE:
 558     retval = "LC_RESPONSE";
 559     break;
 560 #endif
 561 #ifdef LC_ALL
 562   case LC_ALL:
 563     /* This might not make sense but is perhaps better than any other
 564        value.  */
 565     retval = "LC_ALL";
 566     break;
 567 #endif
 568   default:
 569     /* If you have a better idea for a default value let me know.  */
 570     retval = "LC_XXX";
 571   }
 572
 573   return retval;
 574 }
 575
 576 /* Guess value of current locale from value of the environment variables.  */
 577 static const char *
 578 internal_function
 579 guess_category_value (category, categoryname)
 580      int category;
 581      const char *categoryname;
 582 {
 583   const char *retval;
 584
 585   /* The highest priority value is the `LANGUAGE' environment
 586      variable.  This is a GNU extension.  */
 587   retval = getenv ("LANGUAGE");
 588   if (retval != NULL && retval[0] != '\0')
 589     return retval;
 590
 591   /* `LANGUAGE' is not set.  So we have to proceed with the POSIX
 592      methods of looking to `LC_ALL', `LC_xxx', and `LANG'.  On some
 593      systems this can be done by the `setlocale' function itself.  */
 594 #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
 595   return setlocale (category, NULL);
 596 #else
 597   /* Setting of LC_ALL overwrites all other.  */
 598   retval = getenv ("LC_ALL");
 599   if (retval != NULL && retval[0] != '\0')
 600     return retval;
 601
 602   /* Next comes the name of the desired category.  */
 603   retval = getenv (categoryname);
 604   if (retval != NULL && retval[0] != '\0')
 605     return retval;
 606
 607   /* Last possibility is the LANG environment variable.  */
 608   retval = getenv ("LANG");
 609   if (retval != NULL && retval[0] != '\0')
 610     return retval;
 611
 612   /* We use C as the default domain.  POSIX says this is implementation
 613      defined.  */
 614   return "C";
 615 #endif
 616 }
 617
 618 /* @@ begin of epilog @@ */
 619
 620 /* We don't want libintl.a to depend on any other library.  So we
 621    avoid the non-standard function stpcpy.  In GNU C Library this
 622    function is available, though.  Also allow the symbol HAVE_STPCPY
 623    to be defined.  */
 624 #if !_LIBC && !HAVE_STPCPY
 625 static char *
 626 stpcpy (dest, src)
 627      char *dest;
 628      const char *src;
 629 {
 630   while ((*dest++ = *src++) != '\0')
 631     /* Do nothing. */ ;
 632   return dest - 1;
 633 }
 634 #endif
 635
 636
 637 #ifdef _LIBC
 638 /* If we want to free all resources we have to do some work at
 639    program's end.  */
 640 static void __attribute__ ((unused))
 641 free_mem (void)
 642 {
 643   struct binding *runp;
 644
 645   for (runp = _nl_domain_bindings; runp != NULL; runp = runp->next)
 646     {
 647       free (runp->domainname);
 648       if (runp->dirname != _nl_default_dirname)
 649         /* Yes, this is a pointer comparison.  */
 650         free (runp->dirname);
 651     }
 652
 653   if (_nl_current_default_domain != _nl_default_default_domain)
 654     /* Yes, again a pointer comparison.  */
 655     free ((char *) _nl_current_default_domain);
 656 }
 657
 658 text_set_element (__libc_subfreeres, free_mem);
 659 #endif