Home | History | Annotate | Download | only in intl
      1 /* Copyright (C) 1995-1999, 2000-2003 Free Software Foundation, Inc.
      2    Contributed by Ulrich Drepper <drepper (at) gnu.ai.mit.edu>, 1995.
      3 
      4    This program is free software; you can redistribute it and/or modify it
      5    under the terms of the GNU Library General Public License as published
      6    by 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 GNU
     12    Library General Public License for more details.
     13 
     14    You should have received a copy of the GNU Library General Public
     15    License along with this program; if not, write to the Free Software
     16    Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307,
     17    USA.  */
     18 
     19 /* Tell glibc's <string.h> to provide a prototype for stpcpy().
     20    This must come before <config.h> because <config.h> may include
     21    <features.h>, and once <features.h> has been included, it's too late.  */
     22 #ifndef _GNU_SOURCE
     23 # define _GNU_SOURCE	1
     24 #endif
     25 
     26 #ifdef HAVE_CONFIG_H
     27 # include <config.h>
     28 #endif
     29 
     30 #include <string.h>
     31 
     32 #if defined _LIBC || defined HAVE_ARGZ_H
     33 # include <argz.h>
     34 #endif
     35 #include <ctype.h>
     36 #include <sys/types.h>
     37 #include <stdlib.h>
     38 
     39 #include "loadinfo.h"
     40 
     41 /* On some strange systems still no definition of NULL is found.  Sigh!  */
     42 #ifndef NULL
     43 # if defined __STDC__ && __STDC__
     44 #  define NULL ((void *) 0)
     45 # else
     46 #  define NULL 0
     47 # endif
     48 #endif
     49 
     50 /* @@ end of prolog @@ */
     51 
     52 #ifdef _LIBC
     53 /* Rename the non ANSI C functions.  This is required by the standard
     54    because some ANSI C functions will require linking with this object
     55    file and the name space must not be polluted.  */
     56 # ifndef stpcpy
     57 #  define stpcpy(dest, src) __stpcpy(dest, src)
     58 # endif
     59 #else
     60 # ifndef HAVE_STPCPY
     61 #define stpcpy(dest, src) my_stpcpy(dest, src)
     62 static char *stpcpy (char *dest, const char *src);
     63 # endif
     64 #endif
     65 
     66 /* Pathname support.
     67    ISSLASH(C)           tests whether C is a directory separator character.
     68    IS_ABSOLUTE_PATH(P)  tests whether P is an absolute path.  If it is not,
     69                         it may be concatenated to a directory pathname.
     70  */
     71 #if defined _WIN32 || defined __WIN32__ || defined __EMX__ || defined __DJGPP__
     72   /* Win32, OS/2, DOS */
     73 # define ISSLASH(C) ((C) == '/' || (C) == '\\')
     74 # define HAS_DEVICE(P) \
     75     ((((P)[0] >= 'A' && (P)[0] <= 'Z') || ((P)[0] >= 'a' && (P)[0] <= 'z')) \
     76      && (P)[1] == ':')
     77 # define IS_ABSOLUTE_PATH(P) (ISSLASH ((P)[0]) || HAS_DEVICE (P))
     78 #else
     79   /* Unix */
     80 # define ISSLASH(C) ((C) == '/')
     81 # define IS_ABSOLUTE_PATH(P) ISSLASH ((P)[0])
     82 #endif
     83 
     84 /* Define function which are usually not available.  */
     85 
     86 #if !defined _LIBC && !defined HAVE___ARGZ_COUNT
     87 /* Returns the number of strings in ARGZ.  */
     88 static size_t
     89 argz_count__ (const char *argz, size_t len)
     90 {
     91   size_t count = 0;
     92   while (len > 0)
     93     {
     94       size_t part_len = strlen (argz);
     95       argz += part_len + 1;
     96       len -= part_len + 1;
     97       count++;
     98     }
     99   return count;
    100 }
    101 # undef __argz_count
    102 # define __argz_count(argz, len) argz_count__ (argz, len)
    103 #else
    104 # ifdef _LIBC
    105 #  define __argz_count(argz, len) INTUSE(__argz_count) (argz, len)
    106 # endif
    107 #endif	/* !_LIBC && !HAVE___ARGZ_COUNT */
    108 
    109 #if !defined _LIBC && !defined HAVE___ARGZ_STRINGIFY
    110 /* Make '\0' separated arg vector ARGZ printable by converting all the '\0's
    111    except the last into the character SEP.  */
    112 static void
    113 argz_stringify__ (char *argz, size_t len, int sep)
    114 {
    115   while (len > 0)
    116     {
    117       size_t part_len = strlen (argz);
    118       argz += part_len;
    119       len -= part_len + 1;
    120       if (len > 0)
    121 	*argz++ = sep;
    122     }
    123 }
    124 # undef __argz_stringify
    125 # define __argz_stringify(argz, len, sep) argz_stringify__ (argz, len, sep)
    126 #else
    127 # ifdef _LIBC
    128 #  define __argz_stringify(argz, len, sep) \
    129   INTUSE(__argz_stringify) (argz, len, sep)
    130 # endif
    131 #endif	/* !_LIBC && !HAVE___ARGZ_STRINGIFY */
    132 
    133 #if !defined _LIBC && !defined HAVE___ARGZ_NEXT
    134 static char *
    135 argz_next__ (char *argz, size_t argz_len, const char *entry)
    136 {
    137   if (entry)
    138     {
    139       if (entry < argz + argz_len)
    140         entry = strchr (entry, '\0') + 1;
    141 
    142       return entry >= argz + argz_len ? NULL : (char *) entry;
    143     }
    144   else
    145     if (argz_len > 0)
    146       return argz;
    147     else
    148       return 0;
    149 }
    150 # undef __argz_next
    151 # define __argz_next(argz, len, entry) argz_next__ (argz, len, entry)
    152 #endif	/* !_LIBC && !HAVE___ARGZ_NEXT */
    153 
    154 
    155 /* Return number of bits set in X.  */
    156 static inline int
    157 pop (int x)
    158 {
    159   /* We assume that no more than 16 bits are used.  */
    160   x = ((x & ~0x5555) >> 1) + (x & 0x5555);
    161   x = ((x & ~0x3333) >> 2) + (x & 0x3333);
    162   x = ((x >> 4) + x) & 0x0f0f;
    163   x = ((x >> 8) + x) & 0xff;
    164 
    165   return x;
    166 }
    167 
    168 
    169 struct loaded_l10nfile *
    170 _nl_make_l10nflist (struct loaded_l10nfile **l10nfile_list,
    171 		    const char *dirlist, size_t dirlist_len,
    172 		    int mask, const char *language, const char *territory,
    173 		    const char *codeset, const char *normalized_codeset,
    174 		    const char *modifier, const char *special,
    175 		    const char *sponsor, const char *revision,
    176 		    const char *filename, int do_allocate)
    177 {
    178   char *abs_filename;
    179   struct loaded_l10nfile **lastp;
    180   struct loaded_l10nfile *retval;
    181   char *cp;
    182   size_t dirlist_count;
    183   size_t entries;
    184   int cnt;
    185 
    186   /* If LANGUAGE contains an absolute directory specification, we ignore
    187      DIRLIST.  */
    188   if (IS_ABSOLUTE_PATH (language))
    189     dirlist_len = 0;
    190 
    191   /* Allocate room for the full file name.  */
    192   abs_filename = (char *) malloc (dirlist_len
    193 				  + strlen (language)
    194 				  + ((mask & TERRITORY) != 0
    195 				     ? strlen (territory) + 1 : 0)
    196 				  + ((mask & XPG_CODESET) != 0
    197 				     ? strlen (codeset) + 1 : 0)
    198 				  + ((mask & XPG_NORM_CODESET) != 0
    199 				     ? strlen (normalized_codeset) + 1 : 0)
    200 				  + (((mask & XPG_MODIFIER) != 0
    201 				      || (mask & CEN_AUDIENCE) != 0)
    202 				     ? strlen (modifier) + 1 : 0)
    203 				  + ((mask & CEN_SPECIAL) != 0
    204 				     ? strlen (special) + 1 : 0)
    205 				  + (((mask & CEN_SPONSOR) != 0
    206 				      || (mask & CEN_REVISION) != 0)
    207 				     ? (1 + ((mask & CEN_SPONSOR) != 0
    208 					     ? strlen (sponsor) : 0)
    209 					+ ((mask & CEN_REVISION) != 0
    210 					   ? strlen (revision) + 1 : 0)) : 0)
    211 				  + 1 + strlen (filename) + 1);
    212 
    213   if (abs_filename == NULL)
    214     return NULL;
    215 
    216   /* Construct file name.  */
    217   cp = abs_filename;
    218   if (dirlist_len > 0)
    219     {
    220       memcpy (cp, dirlist, dirlist_len);
    221       __argz_stringify (cp, dirlist_len, PATH_SEPARATOR);
    222       cp += dirlist_len;
    223       cp[-1] = '/';
    224     }
    225 
    226   cp = stpcpy (cp, language);
    227 
    228   if ((mask & TERRITORY) != 0)
    229     {
    230       *cp++ = '_';
    231       cp = stpcpy (cp, territory);
    232     }
    233   if ((mask & XPG_CODESET) != 0)
    234     {
    235       *cp++ = '.';
    236       cp = stpcpy (cp, codeset);
    237     }
    238   if ((mask & XPG_NORM_CODESET) != 0)
    239     {
    240       *cp++ = '.';
    241       cp = stpcpy (cp, normalized_codeset);
    242     }
    243   if ((mask & (XPG_MODIFIER | CEN_AUDIENCE)) != 0)
    244     {
    245       /* This component can be part of both syntaces but has different
    246 	 leading characters.  For CEN we use `+', else `@'.  */
    247       *cp++ = (mask & CEN_AUDIENCE) != 0 ? '+' : '@';
    248       cp = stpcpy (cp, modifier);
    249     }
    250   if ((mask & CEN_SPECIAL) != 0)
    251     {
    252       *cp++ = '+';
    253       cp = stpcpy (cp, special);
    254     }
    255   if ((mask & (CEN_SPONSOR | CEN_REVISION)) != 0)
    256     {
    257       *cp++ = ',';
    258       if ((mask & CEN_SPONSOR) != 0)
    259 	cp = stpcpy (cp, sponsor);
    260       if ((mask & CEN_REVISION) != 0)
    261 	{
    262 	  *cp++ = '_';
    263 	  cp = stpcpy (cp, revision);
    264 	}
    265     }
    266 
    267   *cp++ = '/';
    268   stpcpy (cp, filename);
    269 
    270   /* Look in list of already loaded domains whether it is already
    271      available.  */
    272   lastp = l10nfile_list;
    273   for (retval = *l10nfile_list; retval != NULL; retval = retval->next)
    274     if (retval->filename != NULL)
    275       {
    276 	int compare = strcmp (retval->filename, abs_filename);
    277 	if (compare == 0)
    278 	  /* We found it!  */
    279 	  break;
    280 	if (compare < 0)
    281 	  {
    282 	    /* It's not in the list.  */
    283 	    retval = NULL;
    284 	    break;
    285 	  }
    286 
    287 	lastp = &retval->next;
    288       }
    289 
    290   if (retval != NULL || do_allocate == 0)
    291     {
    292       free (abs_filename);
    293       return retval;
    294     }
    295 
    296   dirlist_count = (dirlist_len > 0 ? __argz_count (dirlist, dirlist_len) : 1);
    297 
    298   /* Allocate a new loaded_l10nfile.  */
    299   retval =
    300     (struct loaded_l10nfile *)
    301     malloc (sizeof (*retval)
    302 	    + (((dirlist_count << pop (mask)) + (dirlist_count > 1 ? 1 : 0))
    303 	       * sizeof (struct loaded_l10nfile *)));
    304   if (retval == NULL)
    305     return NULL;
    306 
    307   retval->filename = abs_filename;
    308 
    309   /* We set retval->data to NULL here; it is filled in later.
    310      Setting retval->decided to 1 here means that retval does not
    311      correspond to a real file (dirlist_count > 1) or is not worth
    312      looking up (if an unnormalized codeset was specified).  */
    313   retval->decided = (dirlist_count > 1
    314 		     || ((mask & XPG_CODESET) != 0
    315 			 && (mask & XPG_NORM_CODESET) != 0));
    316   retval->data = NULL;
    317 
    318   retval->next = *lastp;
    319   *lastp = retval;
    320 
    321   entries = 0;
    322   /* Recurse to fill the inheritance list of RETVAL.
    323      If the DIRLIST is a real list (i.e. DIRLIST_COUNT > 1), the RETVAL
    324      entry does not correspond to a real file; retval->filename contains
    325      colons.  In this case we loop across all elements of DIRLIST and
    326      across all bit patterns dominated by MASK.
    327      If the DIRLIST is a single directory or entirely redundant (i.e.
    328      DIRLIST_COUNT == 1), we loop across all bit patterns dominated by
    329      MASK, excluding MASK itself.
    330      In either case, we loop down from MASK to 0.  This has the effect
    331      that the extra bits in the locale name are dropped in this order:
    332      first the modifier, then the territory, then the codeset, then the
    333      normalized_codeset.  */
    334   for (cnt = dirlist_count > 1 ? mask : mask - 1; cnt >= 0; --cnt)
    335     if ((cnt & ~mask) == 0
    336 	&& ((cnt & CEN_SPECIFIC) == 0 || (cnt & XPG_SPECIFIC) == 0)
    337 	&& ((cnt & XPG_CODESET) == 0 || (cnt & XPG_NORM_CODESET) == 0))
    338       {
    339 	if (dirlist_count > 1)
    340 	  {
    341 	    /* Iterate over all elements of the DIRLIST.  */
    342 	    char *dir = NULL;
    343 
    344 	    while ((dir = __argz_next ((char *) dirlist, dirlist_len, dir))
    345 		   != NULL)
    346 	      retval->successor[entries++]
    347 		= _nl_make_l10nflist (l10nfile_list, dir, strlen (dir) + 1,
    348 				      cnt, language, territory, codeset,
    349 				      normalized_codeset, modifier, special,
    350 				      sponsor, revision, filename, 1);
    351 	  }
    352 	else
    353 	  retval->successor[entries++]
    354 	    = _nl_make_l10nflist (l10nfile_list, dirlist, dirlist_len,
    355 				  cnt, language, territory, codeset,
    356 				  normalized_codeset, modifier, special,
    357 				  sponsor, revision, filename, 1);
    358       }
    359   retval->successor[entries] = NULL;
    360 
    361   return retval;
    362 }
    363 
    364 /* Normalize codeset name.  There is no standard for the codeset
    365    names.  Normalization allows the user to use any of the common
    366    names.  The return value is dynamically allocated and has to be
    367    freed by the caller.  */
    368 const char *
    369 _nl_normalize_codeset (const char *codeset, size_t name_len)
    370 {
    371   int len = 0;
    372   int only_digit = 1;
    373   char *retval;
    374   char *wp;
    375   size_t cnt;
    376 
    377   for (cnt = 0; cnt < name_len; ++cnt)
    378     if (isalnum ((unsigned char) codeset[cnt]))
    379       {
    380 	++len;
    381 
    382 	if (isalpha ((unsigned char) codeset[cnt]))
    383 	  only_digit = 0;
    384       }
    385 
    386   retval = (char *) malloc ((only_digit ? 3 : 0) + len + 1);
    387 
    388   if (retval != NULL)
    389     {
    390       if (only_digit)
    391 	wp = stpcpy (retval, "iso");
    392       else
    393 	wp = retval;
    394 
    395       for (cnt = 0; cnt < name_len; ++cnt)
    396 	if (isalpha ((unsigned char) codeset[cnt]))
    397 	  *wp++ = tolower ((unsigned char) codeset[cnt]);
    398 	else if (isdigit ((unsigned char) codeset[cnt]))
    399 	  *wp++ = codeset[cnt];
    400 
    401       *wp = '\0';
    402     }
    403 
    404   return (const char *) retval;
    405 }
    406 
    407 
    408 /* @@ begin of epilog @@ */
    409 
    410 /* We don't want libintl.a to depend on any other library.  So we
    411    avoid the non-standard function stpcpy.  In GNU C Library this
    412    function is available, though.  Also allow the symbol HAVE_STPCPY
    413    to be defined.  */
    414 #if !_LIBC && !HAVE_STPCPY
    415 static char *
    416 stpcpy (char *dest, const char *src)
    417 {
    418   while ((*dest++ = *src++) != '\0')
    419     /* Do nothing. */ ;
    420   return dest - 1;
    421 }
    422 #endif
    423