fc-lang/fc-lang.c

   1 /*
   2  * $RCSId: xc/lib/fontconfig/fc-lang/fc-lang.c,v 1.3 2002/08/22 07:36:43 keithp Exp $
   3  *
   4  * Copyright © 2002 Keith Packard
   5  *
   6  * Permission to use, copy, modify, distribute, and sell this software and its
   7  * documentation for any purpose is hereby granted without fee, provided that
   8  * the above copyright notice appear in all copies and that both that
   9  * copyright notice and this permission notice appear in supporting
  10  * documentation, and that the name of Keith Packard not be used in
  11  * advertising or publicity pertaining to distribution of the software without
  12  * specific, written prior permission.  Keith Packard makes no
  13  * representations about the suitability of this software for any purpose.  It
  14  * is provided "as is" without express or implied warranty.
  15  *
  16  * KEITH PACKARD DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
  17  * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
  18  * EVENT SHALL KEITH PACKARD BE LIABLE FOR ANY SPECIAL, INDIRECT OR
  19  * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
  20  * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
  21  * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
  22  * PERFORMANCE OF THIS SOFTWARE.
  23  */
  24
  25 #include "fcint.h"
  26 #include "fccharset.c"
  27 #include "fcstr.c"
  28
  29 /*
  30  * fc-lang
  31  *
  32  * Read a set of language orthographies and build C declarations for
  33  * charsets which can then be used to identify which languages are
  34  * supported by a given font.  Note that this uses some utilities
  35  * from the fontconfig library, so the necessary file is simply
  36  * included in this compilation.  A couple of extra utility
  37  * functions are also needed in slightly modified form
  38  */
  39
  40 void
  41 FcMemAlloc (int kind, int size)
  42 {
  43 }
  44
  45 void
  46 FcMemFree (int kind, int size)
  47 {
  48 }
  49
  50 FcChar8 *
  51 FcConfigHome (void)
  52 {
  53     return getenv ("HOME");
  54 }
  55
  56 static void
  57 fatal (char *file, int lineno, char *msg)
  58 {
  59     fprintf (stderr, "%s:%d: %s\n", file, lineno, msg);
  60     exit (1);
  61 }
  62
  63 static char *
  64 get_line (FILE *f, char *line, int *lineno)
  65 {
  66     char    *hash;
  67     if (!fgets (line, 1024, f))
  68         return 0;
  69     ++(*lineno);
  70     hash = strchr (line, '#');
  71     if (hash)
  72         *hash = '\0';
  73     if (line[0] == '\0' || line[0] == '\n' || line[0] == '\032' || line[0] == '\r')
  74         return get_line (f, line, lineno);
  75     return line;
  76 }
  77
  78 char    *dir = 0;
  79
  80 static FILE *
  81 scanopen (char *file)
  82 {
  83     FILE    *f;
  84
  85     f = fopen (file, "r");
  86     if (!f && dir)
  87     {
  88         char    path[1024];
  89
  90         strcpy (path, dir);
  91         strcat (path, "/");
  92         strcat (path, file);
  93         f = fopen (path, "r");
  94     }
  95     return f;
  96 }
  97
  98 /*
  99  * build a single charset from a source file
 100  *
 101  * The file format is quite simple, either
 102  * a single hex value or a pair separated with a dash
 103  *
 104  * Comments begin with '#'
 105  */
 106
 107 static FcCharSet *
 108 scan (FILE *f, char *file)
 109 {
 110     FcCharSet   *c = 0;
 111     FcCharSet   *n;
 112     int         start, end, ucs4;
 113     char        line[1024];
 114     int         lineno = 0;
 115
 116     while (get_line (f, line, &lineno))
 117     {
 118         if (!strncmp (line, "include", 7))
 119         {
 120             file = strchr (line, ' ');
 121             while (*file == ' ')
 122                 file++;
 123             end = strlen (file);
 124             if (file[end-1] == '\n')
 125                 file[end-1] = '\0';
 126             f = scanopen (file);
 127             if (!f)
 128                 fatal (file, 0, "can't open");
 129             c = scan (f, file);
 130             fclose (f);
 131             return c;
 132         }
 133         if (strchr (line, '-'))
 134         {
 135             if (sscanf (line, "%x-%x", &start, &end) != 2)
 136                 fatal (file, lineno, "parse error");
 137         }
 138         else
 139         {
 140             if (sscanf (line, "%x", &start) != 1)
 141                 fatal (file, lineno, "parse error");
 142             end = start;
 143         }
 144         if (!c)
 145             c = FcCharSetCreate ();
 146         for (ucs4 = start; ucs4 <= end; ucs4++)
 147         {
 148             if (!FcCharSetAddChar (c, ucs4))
 149                 fatal (file, lineno, "out of memory");
 150         }
 151     }
 152     n = FcCharSetFreeze (c);
 153     FcCharSetDestroy (c);
 154     return n;
 155 }
 156
 157 /*
 158  * Convert a file name into a name suitable for C declarations
 159  */
 160 static char *
 161 get_name (char *file)
 162 {
 163     char    *name;
 164     char    *dot;
 165
 166     dot = strchr (file, '.');
 167     if (!dot)
 168         dot = file + strlen(file);
 169     name = malloc (dot - file + 1);
 170     strncpy (name, file, dot - file);
 171     name[dot-file] = '\0';
 172     return name;
 173 }
 174
 175 /*
 176  * Convert a C name into a language name
 177  */
 178 static char *
 179 get_lang (char *name)
 180 {
 181     char    *lang = malloc (strlen (name) + 1);
 182     char    *l = lang;
 183     char    c;
 184
 185     while ((c = *name++))
 186     {
 187         if (isupper ((int) (unsigned char) c))
 188             c = tolower ((int) (unsigned char) c);
 189         if (c == '_')
 190             c = '-';
 191         if (c == ' ')
 192             continue;
 193         *l++ = c;
 194     }
 195     *l++ = '\0';
 196     return lang;
 197 }
 198
 199 static int compare (const void *a, const void *b)
 200 {
 201     const FcChar8    *const *as = a, *const *bs = b;
 202     return FcStrCmpIgnoreCase (*as, *bs);
 203 }
 204
 205 #define MAX_LANG            1024
 206 #define MAX_LANG_SET_MAP    ((MAX_LANG + 31) / 32)
 207
 208 #define BitSet(map, id)   ((map)[(id)>>5] |= ((FcChar32) 1 << ((id) & 0x1f)))
 209 #define BitGet(map, id)   ((map)[(id)>>5] >> ((id) & 0x1f)) & 1)
 210
 211 int
 212 main (int argc, char **argv)
 213 {
 214     char        *files[MAX_LANG];
 215     FcCharSet   *sets[MAX_LANG];
 216     int         duplicate[MAX_LANG];
 217     int         country[MAX_LANG];
 218     char        *names[MAX_LANG];
 219     char        *langs[MAX_LANG];
 220     FILE        *f;
 221     int         ncountry = 0;
 222     int         i = 0;
 223     FcCharLeaf  **leaves;
 224     int         total_leaves = 0;
 225     int         l, sl, tl;
 226     int         c;
 227     char        line[1024];
 228     FcChar32    map[MAX_LANG_SET_MAP];
 229     int         num_lang_set_map;
 230     int         setRangeStart[26];
 231     int         setRangeEnd[26];
 232     FcChar8     setRangeChar;
 233
 234     while (*++argv)
 235     {
 236         if (!strcmp (*argv, "-d"))
 237         {
 238             dir = *++argv;
 239             continue;
 240         }
 241         if (i == MAX_LANG)
 242             fatal (*argv, 0, "Too many languages");
 243         files[i++] = *argv;
 244     }
 245     files[i] = 0;
 246     qsort (files, i, sizeof (char *), compare);
 247     i = 0;
 248     while (files[i])
 249     {
 250         f = scanopen (files[i]);
 251         if (!f)
 252             fatal (files[i], 0, strerror (errno));
 253         sets[i] = scan (f, files[i]);
 254         names[i] = get_name (files[i]);
 255         langs[i] = get_lang(names[i]);
 256         if (strchr (langs[i], '-'))
 257             country[ncountry++] = i;
 258
 259         total_leaves += sets[i]->num;
 260         i++;
 261         fclose (f);
 262     }
 263     sets[i] = 0;
 264     leaves = malloc (total_leaves * sizeof (FcCharLeaf *));
 265     tl = 0;
 266     /*
 267      * Find unique leaves
 268      */
 269     for (i = 0; sets[i]; i++)
 270     {
 271         for (sl = 0; sl < sets[i]->num; sl++)
 272         {
 273             for (l = 0; l < tl; l++)
 274                 if (leaves[l] == FcCharSetGetLeaf(sets[i], sl))
 275                     break;
 276             if (l == tl)
 277                 leaves[tl++] = FcCharSetGetLeaf(sets[i], sl);
 278         }
 279     }
 280
 281     /*
 282      * Scan the input until the marker is found
 283      */
 284
 285     while (fgets (line, sizeof (line), stdin))
 286     {
 287         if (!strncmp (line, "@@@", 3))
 288             break;
 289         fputs (line, stdout);
 290     }
 291
 292     printf ("/* total size: %d unique leaves: %d */\n\n",
 293             total_leaves, tl);
 294     /*
 295      * Dump leaves
 296      */
 297     printf ("static const FcCharLeaf    leaves[%d] = {\n", tl);
 298     for (l = 0; l < tl; l++)
 299     {
 300         printf ("    { { /* %d */", l);
 301         for (i = 0; i < 256/32; i++)
 302         {
 303             if (i % 4 == 0)
 304                 printf ("\n   ");
 305             printf (" 0x%08x,", leaves[l]->map[i]);
 306         }
 307         printf ("\n    } },\n");
 308     }
 309     printf ("};\n\n");
 310     printf ("#define L(n) ((FcCharLeaf *) &leaves[n])\n\n");
 311
 312     /*
 313      * Find duplicate charsets
 314      */
 315     duplicate[0] = -1;
 316     for (i = 1; sets[i]; i++)
 317     {
 318         int j;
 319
 320         duplicate[i] = -1;
 321         for (j = 0; j < i; j++)
 322             if (sets[j] == sets[i])
 323             {
 324                 duplicate[i] = j;
 325                 break;
 326             }
 327     }
 328
 329     /*
 330      * Find ranges for each letter for faster searching
 331      */
 332     setRangeChar = 'a';
 333     for (i = 0; sets[i]; i++)
 334     {
 335         char    c = names[i][0];
 336
 337         while (setRangeChar <= c && c <= 'z')
 338             setRangeStart[setRangeChar++ - 'a'] = i;
 339     }
 340     for (setRangeChar = 'a'; setRangeChar < 'z'; setRangeChar++)
 341         setRangeEnd[setRangeChar - 'a'] = setRangeStart[setRangeChar+1-'a'] - 1;
 342     setRangeEnd[setRangeChar - 'a'] = i - 1;
 343
 344     /*
 345      * Dump arrays
 346      */
 347     for (i = 0; sets[i]; i++)
 348     {
 349         int n;
 350
 351         if (duplicate[i] >= 0)
 352             continue;
 353         printf ("static const FcCharLeaf *leaves_%s[%d] = {\n",
 354                 names[i], sets[i]->num);
 355         for (n = 0; n < sets[i]->num; n++)
 356         {
 357             if (n % 8 == 0)
 358                 printf ("   ");
 359             for (l = 0; l < tl; l++)
 360                 if (leaves[l] == FcCharSetGetLeaf(sets[i], n))
 361                     break;
 362             if (l == tl)
 363                 fatal (names[i], 0, "can't find leaf");
 364             printf (" L(%3d),", l);
 365             if (n % 8 == 7)
 366                 printf ("\n");
 367         }
 368         if (n % 8 != 0)
 369             printf ("\n");
 370         printf ("};\n\n");
 371
 372
 373         printf ("static const FcChar16 numbers_%s[%d] = {\n",
 374                 names[i], sets[i]->num);
 375         for (n = 0; n < sets[i]->num; n++)
 376         {
 377             if (n % 8 == 0)
 378                 printf ("   ");
 379             printf (" 0x%04x,", FcCharSetGetNumbers(sets[i])[n]);
 380             if (n % 8 == 7)
 381                 printf ("\n");
 382         }
 383         if (n % 8 != 0)
 384             printf ("\n");
 385         printf ("};\n\n");
 386     }
 387     printf ("#undef L\n\n");
 388
 389     /*
 390      * Dump sets
 391      */
 392
 393     printf ("static const FcLangCharSet  fcLangCharSets[] = {\n");
 394     for (i = 0; sets[i]; i++)
 395     {
 396         int     j = duplicate[i];
 397
 398         if (j < 0)
 399             j = i;
 400         printf ("    { (FcChar8 *) \"%s\",\n"
 401                 "      { FC_REF_CONSTANT, %d, FcStorageDynamic, "
 402                 "{ { (FcCharLeaf **) leaves_%s, "
 403                 "(FcChar16 *) numbers_%s } } } },\n",
 404                 langs[i],
 405                 sets[j]->num, names[j], names[j]);
 406     }
 407     printf ("};\n\n");
 408     printf ("#define NUM_LANG_CHAR_SET  %d\n", i);
 409     num_lang_set_map = (i + 31) / 32;
 410     printf ("#define NUM_LANG_SET_MAP   %d\n", num_lang_set_map);
 411     /*
 412      * Dump indices with country codes
 413      */
 414     if (ncountry)
 415     {
 416         int     ncountry_ent = 0;
 417         printf ("\n");
 418         printf ("static const FcChar32 fcLangCountrySets[][NUM_LANG_SET_MAP] = {\n");
 419         for (c = 0; c < ncountry; c++)
 420         {
 421             i = country[c];
 422             if (i >= 0)
 423             {
 424                 int l = strchr (langs[i], '-') - langs[i];
 425                 int d, k;
 426
 427                 for (k = 0; k < num_lang_set_map; k++)
 428                     map[k] = 0;
 429
 430                 BitSet (map, i);
 431                 for (d = c + 1; d < ncountry; d++)
 432                 {
 433                     int j = country[d];
 434                     if (j >= 0 && !strncmp (langs[j], langs[i], l))
 435                     {
 436                         BitSet(map, j);
 437                         country[d] = -1;
 438                     }
 439                 }
 440                 printf ("    {");
 441                 for (k = 0; k < num_lang_set_map; k++)
 442                     printf (" 0x%08x,", map[k]);
 443                 printf (" }, /* %*.*s */\n",
 444                         l, l, langs[i]);
 445                 ++ncountry_ent;
 446             }
 447         }
 448         printf ("};\n\n");
 449         printf ("#define NUM_COUNTRY_SET %d\n", ncountry_ent);
 450     }
 451
 452
 453     /*
 454      * Dump sets start/finish for the fastpath
 455      */
 456     printf ("static const FcLangCharSetRange  fcLangCharSetRanges[] = {\n");
 457     for (setRangeChar = 'a'; setRangeChar <= 'z' ; setRangeChar++)
 458     {
 459         printf ("    { %d, %d }, /* %c */\n",
 460                 setRangeStart[setRangeChar - 'a'],
 461                 setRangeEnd[setRangeChar - 'a'], setRangeChar);
 462     }
 463     printf ("};\n\n");
 464
 465     while (fgets (line, sizeof (line), stdin))
 466         fputs (line, stdout);
 467
 468     fflush (stdout);
 469     exit (ferror (stdout));
 470 }