src/simple-gettext.c

   1 /* simple-gettext.c  - a simplified version of gettext.
   2  * Copyright (C) 1995, 1996, 1997, 1998, 1999 Free Software Foundation, Inc.
   3  *
   4  * This file is part of GnuPG.
   5  *
   6  * GnuPG is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 3 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * GnuPG is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License
  17  * along with this program. If not, see <http://www.gnu.org/licenses/>.
  18  */
  19
  20 /* This is a simplified version of gettext written by Ulrich Drepper.
  21  * It is used for the Win32 version of GnuPG becuase all the overhead
  22  * of gettext is not needed and we have to do some special Win32 stuff.
  23  * I decided that this is far easier than to tweak gettext for the special
  24  * cases (I tried it but it is a lot of code).  wk 15.09.99
  25  */
  26
  27 #include <config.h>
  28 #ifdef USE_SIMPLE_GETTEXT
  29 #ifndef __MINGW32__
  30   #error This file can only be used with MinGW32
  31 #endif
  32
  33 #include <stdio.h>
  34 #include <stdlib.h>
  35 #include <string.h>
  36 #include <ctype.h>
  37 #include <errno.h>
  38 #include <sys/types.h>
  39 #include <sys/stat.h>
  40 #include <windows.h>
  41 #include "w32reg.h"
  42
  43 typedef unsigned int u32; /* That is fine with MingW32 */
  44
  45 typedef unsigned long ulong;
  46
  47 /* The magic number of the GNU message catalog format.  */
  48 #define MAGIC         0x950412de
  49 #define MAGIC_SWAPPED 0xde120495
  50
  51 /* Revision number of the currently used .mo (binary) file format.  */
  52 #define MO_REVISION_NUMBER 0
  53
  54
  55 /* Header for binary .mo file format.  */
  56 struct mo_file_header
  57 {
  58   /* The magic number.  */
  59   u32 magic;
  60   /* The revision number of the file format.  */
  61   u32 revision;
  62   /* The number of strings pairs.  */
  63   u32 nstrings;
  64   /* Offset of table with start offsets of original strings.  */
  65   u32 orig_tab_offset;
  66   /* Offset of table with start offsets of translation strings.  */
  67   u32 trans_tab_offset;
  68   /* Size of hashing table.  */
  69   u32 hash_tab_size;
  70   /* Offset of first hashing entry.  */
  71   u32 hash_tab_offset;
  72 };
  73
  74 struct string_desc
  75 {
  76   /* Length of addressed string.  */
  77   u32 length;
  78   /* Offset of string in file.  */
  79   u32 offset;
  80 };
  81
  82
  83
  84 struct loaded_domain
  85 {
  86   char *data;
  87   int must_swap;
  88   u32 nstrings;
  89   char *mapped;
  90   struct string_desc *orig_tab;
  91   struct string_desc *trans_tab;
  92   u32 hash_size;
  93   u32 *hash_tab;
  94 };
  95
  96
  97 static struct loaded_domain *the_domain;
  98
  99 static __inline__ u32
 100 do_swap_u32( u32 i )
 101 {
 102   return (i << 24) | ((i & 0xff00) << 8) | ((i >> 8) & 0xff00) | (i >> 24);
 103 }
 104
 105 #define SWAPIT(flag, data) ((flag) ? do_swap_u32(data) : (data) )
 106
 107
 108 /* We assume to have `unsigned long int' value with at least 32 bits.  */
 109 #define HASHWORDBITS 32
 110
 111 /* The so called `hashpjw' function by P.J. Weinberger
 112    [see Aho/Sethi/Ullman, COMPILERS: Principles, Techniques and Tools,
 113    1986, 1987 Bell Telephone Laboratories, Inc.]  */
 114
 115 static __inline__ ulong
 116 hash_string( const char *str_param )
 117 {
 118     unsigned long int hval, g;
 119     const char *str = str_param;
 120
 121     hval = 0;
 122     while (*str != '\0')
 123     {
 124         hval <<= 4;
 125         hval += (unsigned long int) *str++;
 126         g = hval & ((unsigned long int) 0xf << (HASHWORDBITS - 4));
 127         if (g != 0)
 128         {
 129           hval ^= g >> (HASHWORDBITS - 8);
 130           hval ^= g;
 131         }
 132     }
 133     return hval;
 134 }
 135
 136
 137 static struct loaded_domain *
 138 load_domain( const char *filename )
 139 {
 140     FILE *fp;
 141     size_t size;
 142     struct stat st;
 143     struct mo_file_header *data = NULL;
 144     struct loaded_domain *domain = NULL;
 145     size_t to_read;
 146     char *read_ptr;
 147
 148     fp = g_fopen( filename, "rb" );
 149     if( !fp )
 150        return NULL; /* can't open the file */
 151     /* we must know about the size of the file */
 152     if( fstat( fileno(fp ), &st )
 153         || (size = (size_t)st.st_size) != st.st_size
 154         || size < sizeof (struct mo_file_header) ) {
 155         fclose( fp );
 156         return NULL;
 157     }
 158
 159     data = malloc( size );
 160     if( !data ) {
 161         fclose( fp );
 162         return NULL; /* out of memory */
 163     }
 164
 165     to_read = size;
 166     read_ptr = (char *) data;
 167     do {
 168         long int nb = fread( read_ptr, 1, to_read, fp );
 169         if( nb < to_read ) {
 170             fclose (fp);
 171             free(data);
 172             return NULL; /* read error */
 173         }
 174         read_ptr += nb;
 175         to_read -= nb;
 176     } while( to_read > 0 );
 177     fclose (fp);
 178
 179     /* Using the magic number we can test whether it really is a message
 180      * catalog file.  */
 181     if( data->magic != MAGIC && data->magic != MAGIC_SWAPPED ) {
 182         /* The magic number is wrong: not a message catalog file.  */
 183         free( data );
 184         return NULL;
 185     }
 186
 187     domain = calloc( 1, sizeof *domain );
 188     if( !domain )  {
 189         free( data );
 190         return NULL;
 191     }
 192     domain->data = (char *) data;
 193     domain->must_swap = data->magic != MAGIC;
 194
 195     /* Fill in the information about the available tables.  */
 196     switch( SWAPIT(domain->must_swap, data->revision) ) {
 197       case 0:
 198         domain->nstrings = SWAPIT(domain->must_swap, data->nstrings);
 199         domain->orig_tab = (struct string_desc *)
 200           ((char *) data + SWAPIT(domain->must_swap, data->orig_tab_offset));
 201         domain->trans_tab = (struct string_desc *)
 202           ((char *) data + SWAPIT(domain->must_swap, data->trans_tab_offset));
 203         domain->hash_size = SWAPIT(domain->must_swap, data->hash_tab_size);
 204         domain->hash_tab = (u32 *)
 205           ((char *) data + SWAPIT(domain->must_swap, data->hash_tab_offset));
 206       break;
 207
 208       default: /* This is an invalid revision.  */
 209         free( data );
 210         free( domain );
 211         return NULL;
 212     }
 213
 214     /* allocate an array to keep track of code page mappings */
 215     domain->mapped = calloc( 1, domain->nstrings );
 216     if( !domain->mapped ) {
 217         free( data );
 218         free( domain );
 219         return NULL;
 220     }
 221
 222     return domain;
 223 }
 224
 225
 226 /****************
 227  * Set the file used for translations.  Pass a NULL to disable
 228  * translation.  A new filename may be set at anytime.
 229  * WARNING: After changing the filename you shoudl not access any data
 230  *          retrieved by gettext().
 231  */
 232 int
 233 set_gettext_file( const char *filename )
 234 {
 235     struct loaded_domain *domain = NULL;
 236
 237     if( filename && *filename ) {
 238         if( filename[0] == '/'
 239            #ifdef HAVE_DRIVE_LETTERS
 240             || ( isalpha(filename[0])
 241                  && filename[1] == ':'
 242                  && (filename[2] == '/' || filename[2] == '\\') )
 243            #endif
 244            ) {
 245             /* absolute path - use it as is */
 246             domain = load_domain( filename );
 247         }
 248         else { /* relative path - append ".mo" and get dir from the environment */
 249             char *buf = NULL;
 250             char *dir;
 251
 252             dir = read_w32_registry_string( NULL,
 253                                             "Control Panel\\Mingw32\\NLS",
 254                                             "MODir" );
 255             if( dir && (buf=malloc(strlen(dir)+strlen(filename)+1+3+1)) ) {
 256                 strcpy(stpcpy(stpcpy(stpcpy( buf, dir),"\\"), filename),".mo");
 257                 domain = load_domain( buf );
 258                 free(buf);
 259             }
 260             free(dir);
 261         }
 262         if( !domain ) {
 263             return -1; }
 264     }
 265     if( the_domain ) {
 266         free( the_domain->data );
 267         free( the_domain->mapped );
 268         free( the_domain );
 269         the_domain = NULL;
 270     }
 271     the_domain = domain;
 272     return NULL;
 273 }
 274
 275
 276 static const char*
 277 get_string( struct loaded_domain *domain, u32 idx )
 278 {
 279     char *p = domain->data + SWAPIT(domain->must_swap,
 280                                     domain->trans_tab[idx].offset);
 281
 282         /* status of domain->mapped[idx] is ignored.
 283          * not sure about the consequences.
 284          * perhaps mapped can entirely be removed?
 285          */
 286
 287         /* we assume, strings are already correctly
 288          * encoded.
 289          */
 290
 291     return (const char*)p;
 292 }
 293
 294
 295
 296 const char *
 297 gettext( const char *msgid )
 298 {
 299     struct loaded_domain *domain;
 300     size_t act = 0;
 301     size_t top, bottom;
 302
 303     if( !(domain = the_domain) ) {
 304         goto not_found;
 305         }
 306
 307     /* Locate the MSGID and its translation.  */
 308     if( domain->hash_size > 2 && domain->hash_tab ) {
 309         /* Use the hashing table.  */
 310         u32 len = strlen (msgid);
 311         u32 hash_val = hash_string (msgid);
 312         u32 idx = hash_val % domain->hash_size;
 313         u32 incr = 1 + (hash_val % (domain->hash_size - 2));
 314         u32 nstr = SWAPIT (domain->must_swap, domain->hash_tab[idx]);
 315
 316         if ( !nstr ) /* Hash table entry is empty.  */
 317             goto not_found;
 318
 319         if( SWAPIT(domain->must_swap,
 320                     domain->orig_tab[nstr - 1].length) == len
 321             && !strcmp( msgid,
 322                        domain->data + SWAPIT(domain->must_swap,
 323                                     domain->orig_tab[nstr - 1].offset)) )
 324             return get_string( domain, nstr - 1 );
 325
 326         for(;;) {
 327             if (idx >= domain->hash_size - incr)
 328                 idx -= domain->hash_size - incr;
 329             else
 330                 idx += incr;
 331
 332             nstr = SWAPIT(domain->must_swap, domain->hash_tab[idx]);
 333             if( !nstr )
 334                 goto not_found; /* Hash table entry is empty.  */
 335
 336             if ( SWAPIT(domain->must_swap,
 337                                 domain->orig_tab[nstr - 1].length) == len
 338                  && !strcmp (msgid,
 339                          domain->data + SWAPIT(domain->must_swap,
 340                                            domain->orig_tab[nstr - 1].offset)))
 341                 return get_string( domain, nstr-1 );
 342         }
 343         /* NOTREACHED */
 344     }
 345
 346     /* Now we try the default method:  binary search in the sorted
 347        array of messages.  */
 348     bottom = 0;
 349     top = domain->nstrings;
 350     while( bottom < top ) {
 351         int cmp_val;
 352
 353         act = (bottom + top) / 2;
 354         cmp_val = strcmp(msgid, domain->data
 355                                + SWAPIT(domain->must_swap,
 356                                         domain->orig_tab[act].offset));
 357         if (cmp_val < 0)
 358             top = act;
 359         else if (cmp_val > 0)
 360             bottom = act + 1;
 361         else
 362             return get_string( domain, act );
 363     }
 364
 365   not_found:
 366     return msgid;
 367 }
 368
 369 #endif /* USE_SIMPLE_GETTEXT */