fc-lang/fc-lang.c

   1 /*
   2  * $XFree86: 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, member of The XFree86 Project, Inc.
   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 /*
  79  * build a single charset from a source file
  80  *
  81  * The file format is quite simple, either
  82  * a single hex value or a pair separated with a dash
  83  *
  84  * Comments begin with '#'
  85  */
  86
  87 static FcCharSet *
  88 scan (FILE *f, char *file)
  89 {
  90     FcCharSet   *c = 0;
  91     FcCharSet   *n;
  92     int         start, end, ucs4;
  93     char        line[1024];
  94     int         lineno = 0;
  95
  96     while (get_line (f, line, &lineno))
  97     {
  98         if (!strncmp (line, "include", 7))
  99         {
 100             file = strchr (line, ' ');
 101             while (*file == ' ')
 102                 file++;
 103             end = strlen (file);
 104             if (file[end-1] == '\n')
 105                 file[end-1] = '\0';
 106             f = fopen (file, "r");
 107             if (!f)
 108                 fatal (file, 0, "can't open");
 109             c = scan (f, file);
 110             fclose (f);
 111             return c;
 112         }
 113         if (strchr (line, '-'))
 114         {
 115             if (sscanf (line, "%x-%x", &start, &end) != 2)
 116                 fatal (file, lineno, "parse error");
 117         }
 118         else
 119         {
 120             if (sscanf (line, "%x", &start) != 1)
 121                 fatal (file, lineno, "parse error");
 122             end = start;
 123         }
 124         if (!c)
 125             c = FcCharSetCreate ();
 126         for (ucs4 = start; ucs4 <= end; ucs4++)
 127         {
 128             if (!FcCharSetAddChar (c, ucs4))
 129                 fatal (file, lineno, "out of memory");
 130         }
 131     }
 132     n = FcCharSetFreeze (c);
 133     FcCharSetDestroy (c);
 134     return n;
 135 }
 136
 137 /*
 138  * Convert a file name into a name suitable for C declarations
 139  */
 140 static char *
 141 get_name (char *file)
 142 {
 143     char    *name;
 144     char    *dot;
 145
 146     dot = strchr (file, '.');
 147     if (!dot)
 148         dot = file + strlen(file);
 149     name = malloc (dot - file + 1);
 150     strncpy (name, file, dot - file);
 151     name[dot-file] = '\0';
 152     return name;
 153 }
 154
 155 /*
 156  * Convert a C name into a language name
 157  */
 158 static char *
 159 get_lang (char *name)
 160 {
 161     char    *lang = malloc (strlen (name) + 1);
 162     char    *l = lang;
 163     char    c;
 164
 165     while ((c = *name++))
 166     {
 167         if (isupper (c))
 168             c = tolower (c);
 169         if (c == '_')
 170             c = '-';
 171         if (c == ' ')
 172             continue;
 173         *l++ = c;
 174     }
 175     *l++ = '\0';
 176     return lang;
 177 }
 178
 179 static int compare (const void *a, const void *b)
 180 {
 181     const FcChar8    *const *as = a, *const *bs = b;
 182     return FcStrCmpIgnoreCase (*as, *bs);
 183 }
 184
 185 #define MAX_LANG            1024
 186 #define MAX_LANG_SET_MAP    ((MAX_LANG + 31) / 32)
 187
 188 #define BitSet(map, id)   ((map)[(id)>>5] |= ((FcChar32) 1 << ((id) & 0x1f)))
 189 #define BitGet(map, id)   ((map)[(id)>>5] >> ((id) & 0x1f)) & 1)
 190
 191 int
 192 main (int argc, char **argv)
 193 {
 194     char        *files[MAX_LANG];
 195     FcCharSet   *sets[MAX_LANG];
 196     int         duplicate[MAX_LANG];
 197     int         country[MAX_LANG];
 198     char        *names[MAX_LANG];
 199     char        *langs[MAX_LANG];
 200     FILE        *f;
 201     int         ncountry = 0;
 202     int         i = 0;
 203     FcCharLeaf  **leaves, **sleaves;
 204     int         total_leaves = 0;
 205     int         l, sl, tl;
 206     int         c;
 207     char        line[1024];
 208     FcChar32    map[MAX_LANG_SET_MAP];
 209     int         num_lang_set_map;
 210
 211     while (*++argv)
 212     {
 213         if (i == MAX_LANG)
 214             fatal (*argv, 0, "Too many languages");
 215         files[i++] = *argv;
 216     }
 217     files[i] = 0;
 218     qsort (files, i, sizeof (char *), compare);
 219     i = 0;
 220     while (files[i])
 221     {
 222         f = fopen (files[i], "r");
 223         if (!f)
 224             fatal (files[i], 0, strerror (errno));
 225         sets[i] = scan (f, files[i]);
 226         names[i] = get_name (files[i]);
 227         langs[i] = get_lang(names[i]);
 228         if (strchr (langs[i], '-'))
 229             country[ncountry++] = i;
 230
 231         total_leaves += sets[i]->num;
 232         i++;
 233         fclose (f);
 234     }
 235     sets[i] = 0;
 236     leaves = malloc (total_leaves * sizeof (FcCharLeaf *));
 237     tl = 0;
 238     /*
 239      * Find unique leaves
 240      */
 241     for (i = 0; sets[i]; i++)
 242     {
 243         sleaves = sets[i]->leaves;
 244         for (sl = 0; sl < sets[i]->num; sl++)
 245         {
 246             for (l = 0; l < tl; l++)
 247                 if (leaves[l] == sleaves[sl])
 248                     break;
 249             if (l == tl)
 250                 leaves[tl++] = sleaves[sl];
 251         }
 252     }
 253
 254     /*
 255      * Scan the input until the marker is found
 256      */
 257
 258     while (fgets (line, sizeof (line), stdin))
 259     {
 260         if (!strncmp (line, "@@@", 3))
 261             break;
 262         fputs (line, stdout);
 263     }
 264
 265     printf ("/* total size: %d unique leaves: %d */\n\n",
 266             total_leaves, tl);
 267     /*
 268      * Dump leaves
 269      */
 270     printf ("static const FcCharLeaf    leaves[%d] = {\n", tl);
 271     for (l = 0; l < tl; l++)
 272     {
 273         printf ("    { { /* %d */", l);
 274         for (i = 0; i < 256/32; i++)
 275         {
 276             if (i % 4 == 0)
 277                 printf ("\n   ");
 278             printf (" 0x%08x,", leaves[l]->map[i]);
 279         }
 280         printf ("\n    } },\n");
 281     }
 282     printf ("};\n\n");
 283     printf ("#define L(n) ((FcCharLeaf *) &leaves[n])\n\n");
 284
 285     /*
 286      * Find duplicate charsets
 287      */
 288     duplicate[0] = -1;
 289     for (i = 1; sets[i]; i++)
 290     {
 291         int j;
 292
 293         duplicate[i] = -1;
 294         for (j = 0; j < i; j++)
 295             if (sets[j] == sets[i])
 296             {
 297                 duplicate[i] = j;
 298                 break;
 299             }
 300     }
 301
 302     /*
 303      * Dump arrays
 304      */
 305     for (i = 0; sets[i]; i++)
 306     {
 307         int n;
 308
 309         if (duplicate[i] >= 0)
 310             continue;
 311         printf ("static const FcCharLeaf *leaves_%s[%d] = {\n",
 312                 names[i], sets[i]->num);
 313         for (n = 0; n < sets[i]->num; n++)
 314         {
 315             if (n % 8 == 0)
 316                 printf ("   ");
 317             for (l = 0; l < tl; l++)
 318                 if (leaves[l] == sets[i]->leaves[n])
 319                     break;
 320             if (l == tl)
 321                 fatal (names[i], 0, "can't find leaf");
 322             printf (" L(%3d),", l);
 323             if (n % 8 == 7)
 324                 printf ("\n");
 325         }
 326         if (n % 8 != 0)
 327             printf ("\n");
 328         printf ("};\n\n");
 329
 330
 331         printf ("static const FcChar16 numbers_%s[%d] = {\n",
 332                 names[i], sets[i]->num);
 333         for (n = 0; n < sets[i]->num; n++)
 334         {
 335             if (n % 8 == 0)
 336                 printf ("   ");
 337             printf (" 0x%04x,", sets[i]->numbers[n]);
 338             if (n % 8 == 7)
 339                 printf ("\n");
 340         }
 341         if (n % 8 != 0)
 342             printf ("\n");
 343         printf ("};\n\n");
 344     }
 345     printf ("#undef L\n\n");
 346     /*
 347      * Dump sets
 348      */
 349     printf ("static const FcLangCharSet  fcLangCharSets[] = {\n");
 350     for (i = 0; sets[i]; i++)
 351     {
 352         int     j = duplicate[i];
 353         if (j < 0)
 354             j = i;
 355         printf ("    { (FcChar8 *) \"%s\",\n"
 356                 "      { FC_REF_CONSTANT, %d, "
 357                 "(FcCharLeaf **) leaves_%s, "
 358                 "(FcChar16 *) numbers_%s } },\n",
 359                 langs[i],
 360                 sets[j]->num, names[j], names[j]);
 361     }
 362     printf ("};\n\n");
 363     printf ("#define NUM_LANG_CHAR_SET  %d\n", i);
 364     num_lang_set_map = (i + 31) / 32;
 365     printf ("#define NUM_LANG_SET_MAP   %d\n", num_lang_set_map);
 366     /*
 367      * Dump indices with country codes
 368      */
 369     if (ncountry)
 370     {
 371         int     ncountry_ent = 0;
 372         printf ("\n");
 373         printf ("static const FcChar32 fcLangCountrySets[][NUM_LANG_SET_MAP] = {\n");
 374         for (c = 0; c < ncountry; c++)
 375         {
 376             i = country[c];
 377             if (i >= 0)
 378             {
 379                 int l = strchr (langs[i], '-') - langs[i];
 380                 int d, k;
 381
 382                 for (k = 0; k < num_lang_set_map; k++)
 383                     map[k] = 0;
 384
 385                 BitSet (map, i);
 386                 for (d = c + 1; d < ncountry; d++)
 387                 {
 388                     int j = country[d];
 389                     if (j >= 0 && !strncmp (langs[j], langs[i], l))
 390                     {
 391                         BitSet(map, j);
 392                         country[d] = -1;
 393                     }
 394                 }
 395                 printf ("    {");
 396                 for (k = 0; k < num_lang_set_map; k++)
 397                     printf (" 0x%08x,", map[k]);
 398                 printf (" }, /* %*.*s */\n",
 399                         l, l, langs[i]);
 400                 ++ncountry_ent;
 401             }
 402         }
 403         printf ("};\n\n");
 404         printf ("#define NUM_COUNTRY_SET %d\n", ncountry_ent);
 405     }
 406
 407     while (fgets (line, sizeof (line), stdin))
 408         fputs (line, stdout);
 409
 410     fflush (stdout);
 411     exit (ferror (stdout));
 412 }