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
  27 /*
  28  * fc-lang
  29  *
  30  * Read a set of language orthographies and build C declarations for
  31  * charsets which can then be used to identify which languages are
  32  * supported by a given font.  Note that it would be nice if
  33  * this could be done while compiling the library, but this
  34  * code uses a number of routines from the library.  It's
  35  * expediant to just ship the pre-built version along with the
  36  * source orthographies.
  37  */
  38
  39 static void
  40 fatal (char *file, int lineno, char *msg)
  41 {
  42     fprintf (stderr, "%s:%d: %s\n", file, lineno, msg);
  43     exit (1);
  44 }
  45
  46 static char *
  47 get_line (FILE *f, char *line, int *lineno)
  48 {
  49     char    *hash;
  50     if (!fgets (line, 1024, f))
  51         return 0;
  52     ++(*lineno);
  53     hash = strchr (line, '#');
  54     if (hash)
  55         *hash = '\0';
  56     if (line[0] == '\0' || line[0] == '\n' || line[0] == '\032' || line[0] == '\r')
  57         return get_line (f, line, lineno);
  58     return line;
  59 }
  60
  61 /*
  62  * build a single charset from a source file
  63  *
  64  * The file format is quite simple, either
  65  * a single hex value or a pair separated with a dash
  66  *
  67  * Comments begin with '#'
  68  */
  69
  70 static FcCharSet *
  71 scan (FILE *f, char *file)
  72 {
  73     FcCharSet   *c = 0;
  74     FcCharSet   *n;
  75     int         start, end, ucs4;
  76     char        line[1024];
  77     int         lineno = 0;
  78
  79     while (get_line (f, line, &lineno))
  80     {
  81         if (!strncmp (line, "include", 7))
  82         {
  83             file = strchr (line, ' ');
  84             while (*file == ' ')
  85                 file++;
  86             end = strlen (file);
  87             if (file[end-1] == '\n')
  88                 file[end-1] = '\0';
  89             f = fopen (file, "r");
  90             if (!f)
  91                 fatal (file, 0, "can't open");
  92             c = scan (f, file);
  93             fclose (f);
  94             return c;
  95         }
  96         if (strchr (line, '-'))
  97         {
  98             if (sscanf (line, "%x-%x", &start, &end) != 2)
  99                 fatal (file, lineno, "parse error");
 100         }
 101         else
 102         {
 103             if (sscanf (line, "%x", &start) != 1)
 104                 fatal (file, lineno, "parse error");
 105             end = start;
 106         }
 107         if (!c)
 108             c = FcCharSetCreate ();
 109         for (ucs4 = start; ucs4 <= end; ucs4++)
 110         {
 111             if (!FcCharSetAddChar (c, ucs4))
 112                 fatal (file, lineno, "out of memory");
 113         }
 114     }
 115     n = FcCharSetFreeze (c);
 116     FcCharSetDestroy (c);
 117     return n;
 118 }
 119
 120 /*
 121  * Convert a file name into a name suitable for C declarations
 122  */
 123 static char *
 124 get_name (char *file)
 125 {
 126     char    *name;
 127     char    *dot;
 128
 129     dot = strchr (file, '.');
 130     if (!dot)
 131         dot = file + strlen(file);
 132     name = malloc (dot - file + 1);
 133     strncpy (name, file, dot - file);
 134     name[dot-file] = '\0';
 135     return name;
 136 }
 137
 138 /*
 139  * Convert a C name into a language name
 140  */
 141 static char *
 142 get_lang (char *name)
 143 {
 144     char    *lang = malloc (strlen (name) + 1);
 145     char    *l = lang;
 146     char    c;
 147
 148     while ((c = *name++))
 149     {
 150         if (isupper (c))
 151             c = tolower (c);
 152         if (c == '_')
 153             c = '-';
 154         if (c == ' ')
 155             continue;
 156         *l++ = c;
 157     }
 158     *l++ = '\0';
 159     return lang;
 160 }
 161
 162 static int compare (const void *a, const void *b)
 163 {
 164     const FcChar8    *const *as = a, *const *bs = b;
 165     return FcStrCmpIgnoreCase (*as, *bs);
 166 }
 167
 168 #define MAX_LANG            1024
 169 #define MAX_LANG_SET_MAP    ((MAX_LANG + 31) / 32)
 170
 171 #define BitSet(map, id)   ((map)[(id)>>5] |= ((FcChar32) 1 << ((id) & 0x1f)))
 172 #define BitGet(map, id)   ((map)[(id)>>5] >> ((id) & 0x1f)) & 1)
 173
 174 int
 175 main (int argc, char **argv)
 176 {
 177     char        *files[MAX_LANG];
 178     FcCharSet   *sets[MAX_LANG];
 179     int         duplicate[MAX_LANG];
 180     int         country[MAX_LANG];
 181     char        *names[MAX_LANG];
 182     char        *langs[MAX_LANG];
 183     FILE        *f;
 184     int         ncountry = 0;
 185     int         i = 0;
 186     FcCharLeaf  **leaves, **sleaves;
 187     int         total_leaves = 0;
 188     int         l, sl, tl;
 189     int         c;
 190     char        line[1024];
 191     FcChar32    map[MAX_LANG_SET_MAP];
 192     int         num_lang_set_map;
 193
 194     while (*++argv)
 195     {
 196         if (i == MAX_LANG)
 197             fatal (*argv, 0, "Too many languages");
 198         files[i++] = *argv;
 199     }
 200     files[i] = 0;
 201     qsort (files, i, sizeof (char *), compare);
 202     i = 0;
 203     while (files[i])
 204     {
 205         f = fopen (files[i], "r");
 206         if (!f)
 207             fatal (files[i], 0, strerror (errno));
 208         sets[i] = scan (f, files[i]);
 209         names[i] = get_name (files[i]);
 210         langs[i] = get_lang(names[i]);
 211         if (strchr (langs[i], '-'))
 212             country[ncountry++] = i;
 213
 214         total_leaves += sets[i]->num;
 215         i++;
 216         fclose (f);
 217     }
 218     sets[i] = 0;
 219     leaves = malloc (total_leaves * sizeof (FcCharLeaf *));
 220     tl = 0;
 221     /*
 222      * Find unique leaves
 223      */
 224     for (i = 0; sets[i]; i++)
 225     {
 226         sleaves = sets[i]->leaves;
 227         for (sl = 0; sl < sets[i]->num; sl++)
 228         {
 229             for (l = 0; l < tl; l++)
 230                 if (leaves[l] == sleaves[sl])
 231                     break;
 232             if (l == tl)
 233                 leaves[tl++] = sleaves[sl];
 234         }
 235     }
 236
 237     /*
 238      * Scan the input until the marker is found
 239      */
 240
 241     while (fgets (line, sizeof (line), stdin))
 242     {
 243         if (!strncmp (line, "@@@", 3))
 244             break;
 245         fputs (line, stdout);
 246     }
 247
 248     printf ("/* total size: %d unique leaves: %d */\n\n",
 249             total_leaves, tl);
 250     /*
 251      * Dump leaves
 252      */
 253     printf ("static const FcCharLeaf    leaves[%d] = {\n", tl);
 254     for (l = 0; l < tl; l++)
 255     {
 256         printf ("    { { /* %d */", l);
 257         for (i = 0; i < 256/32; i++)
 258         {
 259             if (i % 4 == 0)
 260                 printf ("\n   ");
 261             printf (" 0x%08x,", leaves[l]->map[i]);
 262         }
 263         printf ("\n    } },\n");
 264     }
 265     printf ("};\n\n");
 266     printf ("#define L(n) ((FcCharLeaf *) &leaves[n])\n\n");
 267
 268     /*
 269      * Find duplicate charsets
 270      */
 271     duplicate[0] = -1;
 272     for (i = 1; sets[i]; i++)
 273     {
 274         int j;
 275
 276         duplicate[i] = -1;
 277         for (j = 0; j < i; j++)
 278             if (sets[j] == sets[i])
 279             {
 280                 duplicate[i] = j;
 281                 break;
 282             }
 283     }
 284
 285     /*
 286      * Dump arrays
 287      */
 288     for (i = 0; sets[i]; i++)
 289     {
 290         int n;
 291
 292         if (duplicate[i] >= 0)
 293             continue;
 294         printf ("static const FcCharLeaf *leaves_%s[%d] = {\n",
 295                 names[i], sets[i]->num);
 296         for (n = 0; n < sets[i]->num; n++)
 297         {
 298             if (n % 8 == 0)
 299                 printf ("   ");
 300             for (l = 0; l < tl; l++)
 301                 if (leaves[l] == sets[i]->leaves[n])
 302                     break;
 303             if (l == tl)
 304                 fatal (names[i], 0, "can't find leaf");
 305             printf (" L(%3d),", l);
 306             if (n % 8 == 7)
 307                 printf ("\n");
 308         }
 309         if (n % 8 != 0)
 310             printf ("\n");
 311         printf ("};\n\n");
 312
 313
 314         printf ("static const FcChar16 numbers_%s[%d] = {\n",
 315                 names[i], sets[i]->num);
 316         for (n = 0; n < sets[i]->num; n++)
 317         {
 318             if (n % 8 == 0)
 319                 printf ("   ");
 320             printf (" 0x%04x,", sets[i]->numbers[n]);
 321             if (n % 8 == 7)
 322                 printf ("\n");
 323         }
 324         if (n % 8 != 0)
 325             printf ("\n");
 326         printf ("};\n\n");
 327     }
 328     printf ("#undef L\n\n");
 329     /*
 330      * Dump sets
 331      */
 332     printf ("static const FcLangCharSet  fcLangCharSets[] = {\n");
 333     for (i = 0; sets[i]; i++)
 334     {
 335         int     j = duplicate[i];
 336         if (j < 0)
 337             j = i;
 338         printf ("    { (FcChar8 *) \"%s\",\n"
 339                 "      { FC_REF_CONSTANT, %d, "
 340                 "(FcCharLeaf **) leaves_%s, "
 341                 "(FcChar16 *) numbers_%s } },\n",
 342                 langs[i],
 343                 sets[j]->num, names[j], names[j]);
 344     }
 345     printf ("};\n\n");
 346     printf ("#define NUM_LANG_CHAR_SET  %d\n", i);
 347     num_lang_set_map = (i + 31) / 32;
 348     printf ("#define NUM_LANG_SET_MAP   %d\n", num_lang_set_map);
 349     /*
 350      * Dump indices with country codes
 351      */
 352     if (ncountry)
 353     {
 354         int     ncountry_ent = 0;
 355         printf ("\n");
 356         printf ("static const FcChar32 fcLangCountrySets[][NUM_LANG_SET_MAP] = {\n");
 357         for (c = 0; c < ncountry; c++)
 358         {
 359             i = country[c];
 360             if (i >= 0)
 361             {
 362                 int l = strchr (langs[i], '-') - langs[i];
 363                 int d, k;
 364
 365                 for (k = 0; k < num_lang_set_map; k++)
 366                     map[k] = 0;
 367
 368                 BitSet (map, i);
 369                 for (d = c + 1; d < ncountry; d++)
 370                 {
 371                     int j = country[d];
 372                     if (j >= 0 && !strncmp (langs[j], langs[i], l))
 373                     {
 374                         BitSet(map, j);
 375                         country[d] = -1;
 376                     }
 377                 }
 378                 printf ("    {");
 379                 for (k = 0; k < num_lang_set_map; k++)
 380                     printf (" 0x%08x,", map[k]);
 381                 printf (" }, /* %*.*s */\n",
 382                         l, l, langs[i]);
 383                 ++ncountry_ent;
 384             }
 385         }
 386         printf ("};\n\n");
 387         printf ("#define NUM_COUNTRY_SET %d\n", ncountry_ent);
 388     }
 389
 390     while (fgets (line, sizeof (line), stdin))
 391         fputs (line, stdout);
 392
 393     fflush (stdout);
 394     exit (ferror (stdout));
 395 }