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 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, **sleaves;
 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         sleaves = sets[i]->leaves;
 272         for (sl = 0; sl < sets[i]->num; sl++)
 273         {
 274             for (l = 0; l < tl; l++)
 275                 if (leaves[l] == sleaves[sl])
 276                     break;
 277             if (l == tl)
 278                 leaves[tl++] = sleaves[sl];
 279         }
 280     }
 281
 282     /*
 283      * Scan the input until the marker is found
 284      */
 285
 286     while (fgets (line, sizeof (line), stdin))
 287     {
 288         if (!strncmp (line, "@@@", 3))
 289             break;
 290         fputs (line, stdout);
 291     }
 292
 293     printf ("/* total size: %d unique leaves: %d */\n\n",
 294             total_leaves, tl);
 295     /*
 296      * Dump leaves
 297      */
 298     printf ("static const FcCharLeaf    leaves[%d] = {\n", tl);
 299     for (l = 0; l < tl; l++)
 300     {
 301         printf ("    { { /* %d */", l);
 302         for (i = 0; i < 256/32; i++)
 303         {
 304             if (i % 4 == 0)
 305                 printf ("\n   ");
 306             printf (" 0x%08x,", leaves[l]->map[i]);
 307         }
 308         printf ("\n    } },\n");
 309     }
 310     printf ("};\n\n");
 311     printf ("#define L(n) ((FcCharLeaf *) &leaves[n])\n\n");
 312
 313     /*
 314      * Find duplicate charsets
 315      */
 316     duplicate[0] = -1;
 317     for (i = 1; sets[i]; i++)
 318     {
 319         int j;
 320
 321         duplicate[i] = -1;
 322         for (j = 0; j < i; j++)
 323             if (sets[j] == sets[i])
 324             {
 325                 duplicate[i] = j;
 326                 break;
 327             }
 328     }
 329
 330     /*
 331      * Find ranges for each letter for faster searching
 332      */
 333     setRangeChar = 'a';
 334     for (i = 0; sets[i]; i++)
 335     {
 336         char    c = names[i][0];
 337
 338         while (setRangeChar <= c && c <= 'z')
 339             setRangeStart[setRangeChar++ - 'a'] = i;
 340     }
 341     for (setRangeChar = 'a'; setRangeChar < 'z'; setRangeChar++)
 342         setRangeEnd[setRangeChar - 'a'] = setRangeStart[setRangeChar+1-'a'] - 1;
 343     setRangeEnd[setRangeChar - 'a'] = i - 1;
 344
 345     /*
 346      * Dump arrays
 347      */
 348     for (i = 0; sets[i]; i++)
 349     {
 350         int n;
 351
 352         if (duplicate[i] >= 0)
 353             continue;
 354         printf ("static const FcCharLeaf *leaves_%s[%d] = {\n",
 355                 names[i], sets[i]->num);
 356         for (n = 0; n < sets[i]->num; n++)
 357         {
 358             if (n % 8 == 0)
 359                 printf ("   ");
 360             for (l = 0; l < tl; l++)
 361                 if (leaves[l] == sets[i]->leaves[n])
 362                     break;
 363             if (l == tl)
 364                 fatal (names[i], 0, "can't find leaf");
 365             printf (" L(%3d),", l);
 366             if (n % 8 == 7)
 367                 printf ("\n");
 368         }
 369         if (n % 8 != 0)
 370             printf ("\n");
 371         printf ("};\n\n");
 372
 373
 374         printf ("static const FcChar16 numbers_%s[%d] = {\n",
 375                 names[i], sets[i]->num);
 376         for (n = 0; n < sets[i]->num; n++)
 377         {
 378             if (n % 8 == 0)
 379                 printf ("   ");
 380             printf (" 0x%04x,", sets[i]->numbers[n]);
 381             if (n % 8 == 7)
 382                 printf ("\n");
 383         }
 384         if (n % 8 != 0)
 385             printf ("\n");
 386         printf ("};\n\n");
 387     }
 388     printf ("#undef L\n\n");
 389
 390     /*
 391      * Dump sets
 392      */
 393
 394     printf ("static const FcLangCharSet  fcLangCharSets[] = {\n");
 395     for (i = 0; sets[i]; i++)
 396     {
 397         int     j = duplicate[i];
 398
 399         if (j < 0)
 400             j = i;
 401         printf ("    { (FcChar8 *) \"%s\",\n"
 402                 "      { FC_REF_CONSTANT, %d, "
 403                 "(FcCharLeaf **) leaves_%s, "
 404                 "(FcChar16 *) numbers_%s } },\n",
 405                 langs[i],
 406                 sets[j]->num, names[j], names[j]);
 407     }
 408     printf ("};\n\n");
 409     printf ("#define NUM_LANG_CHAR_SET  %d\n", i);
 410     num_lang_set_map = (i + 31) / 32;
 411     printf ("#define NUM_LANG_SET_MAP   %d\n", num_lang_set_map);
 412     /*
 413      * Dump indices with country codes
 414      */
 415     if (ncountry)
 416     {
 417         int     ncountry_ent = 0;
 418         printf ("\n");
 419         printf ("static const FcChar32 fcLangCountrySets[][NUM_LANG_SET_MAP] = {\n");
 420         for (c = 0; c < ncountry; c++)
 421         {
 422             i = country[c];
 423             if (i >= 0)
 424             {
 425                 int l = strchr (langs[i], '-') - langs[i];
 426                 int d, k;
 427
 428                 for (k = 0; k < num_lang_set_map; k++)
 429                     map[k] = 0;
 430
 431                 BitSet (map, i);
 432                 for (d = c + 1; d < ncountry; d++)
 433                 {
 434                     int j = country[d];
 435                     if (j >= 0 && !strncmp (langs[j], langs[i], l))
 436                     {
 437                         BitSet(map, j);
 438                         country[d] = -1;
 439                     }
 440                 }
 441                 printf ("    {");
 442                 for (k = 0; k < num_lang_set_map; k++)
 443                     printf (" 0x%08x,", map[k]);
 444                 printf (" }, /* %*.*s */\n",
 445                         l, l, langs[i]);
 446                 ++ncountry_ent;
 447             }
 448         }
 449         printf ("};\n\n");
 450         printf ("#define NUM_COUNTRY_SET %d\n", ncountry_ent);
 451     }
 452
 453
 454     /*
 455      * Dump sets start/finish for the fastpath
 456      */
 457     printf ("static const FcLangCharSetRange  fcLangCharSetRanges[] = {\n");
 458     for (setRangeChar = 'a'; setRangeChar <= 'z' ; setRangeChar++)
 459     {
 460         printf ("    { %d, %d }, /* %c */\n",
 461                 setRangeStart[setRangeChar - 'a'],
 462                 setRangeEnd[setRangeChar - 'a'], setRangeChar);
 463     }
 464     printf ("};\n\n");
 465
 466     while (fgets (line, sizeof (line), stdin))
 467         fputs (line, stdout);
 468
 469     fflush (stdout);
 470     exit (ferror (stdout));
 471 }