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 static void
  51 fatal (char *file, int lineno, char *msg)
  52 {
  53     fprintf (stderr, "%s:%d: %s\n", file, lineno, msg);
  54     exit (1);
  55 }
  56
  57 static char *
  58 get_line (FILE *f, char *line, int *lineno)
  59 {
  60     char    *hash;
  61     if (!fgets (line, 1024, f))
  62         return 0;
  63     ++(*lineno);
  64     hash = strchr (line, '#');
  65     if (hash)
  66         *hash = '\0';
  67     if (line[0] == '\0' || line[0] == '\n' || line[0] == '\032' || line[0] == '\r')
  68         return get_line (f, line, lineno);
  69     return line;
  70 }
  71
  72 /*
  73  * build a single charset from a source file
  74  *
  75  * The file format is quite simple, either
  76  * a single hex value or a pair separated with a dash
  77  *
  78  * Comments begin with '#'
  79  */
  80
  81 static FcCharSet *
  82 scan (FILE *f, char *file)
  83 {
  84     FcCharSet   *c = 0;
  85     FcCharSet   *n;
  86     int         start, end, ucs4;
  87     char        line[1024];
  88     int         lineno = 0;
  89
  90     while (get_line (f, line, &lineno))
  91     {
  92         if (!strncmp (line, "include", 7))
  93         {
  94             file = strchr (line, ' ');
  95             while (*file == ' ')
  96                 file++;
  97             end = strlen (file);
  98             if (file[end-1] == '\n')
  99                 file[end-1] = '\0';
 100             f = fopen (file, "r");
 101             if (!f)
 102                 fatal (file, 0, "can't open");
 103             c = scan (f, file);
 104             fclose (f);
 105             return c;
 106         }
 107         if (strchr (line, '-'))
 108         {
 109             if (sscanf (line, "%x-%x", &start, &end) != 2)
 110                 fatal (file, lineno, "parse error");
 111         }
 112         else
 113         {
 114             if (sscanf (line, "%x", &start) != 1)
 115                 fatal (file, lineno, "parse error");
 116             end = start;
 117         }
 118         if (!c)
 119             c = FcCharSetCreate ();
 120         for (ucs4 = start; ucs4 <= end; ucs4++)
 121         {
 122             if (!FcCharSetAddChar (c, ucs4))
 123                 fatal (file, lineno, "out of memory");
 124         }
 125     }
 126     n = FcCharSetFreeze (c);
 127     FcCharSetDestroy (c);
 128     return n;
 129 }
 130
 131 /*
 132  * Convert a file name into a name suitable for C declarations
 133  */
 134 static char *
 135 get_name (char *file)
 136 {
 137     char    *name;
 138     char    *dot;
 139
 140     dot = strchr (file, '.');
 141     if (!dot)
 142         dot = file + strlen(file);
 143     name = malloc (dot - file + 1);
 144     strncpy (name, file, dot - file);
 145     name[dot-file] = '\0';
 146     return name;
 147 }
 148
 149 /*
 150  * Convert a C name into a language name
 151  */
 152 static char *
 153 get_lang (char *name)
 154 {
 155     char    *lang = malloc (strlen (name) + 1);
 156     char    *l = lang;
 157     char    c;
 158
 159     while ((c = *name++))
 160     {
 161         if (isupper (c))
 162             c = tolower (c);
 163         if (c == '_')
 164             c = '-';
 165         if (c == ' ')
 166             continue;
 167         *l++ = c;
 168     }
 169     *l++ = '\0';
 170     return lang;
 171 }
 172
 173 static int compare (const void *a, const void *b)
 174 {
 175     const FcChar8    *const *as = a, *const *bs = b;
 176     return FcStrCmpIgnoreCase (*as, *bs);
 177 }
 178
 179 #define MAX_LANG            1024
 180 #define MAX_LANG_SET_MAP    ((MAX_LANG + 31) / 32)
 181
 182 #define BitSet(map, id)   ((map)[(id)>>5] |= ((FcChar32) 1 << ((id) & 0x1f)))
 183 #define BitGet(map, id)   ((map)[(id)>>5] >> ((id) & 0x1f)) & 1)
 184
 185 int
 186 main (int argc, char **argv)
 187 {
 188     char        *files[MAX_LANG];
 189     FcCharSet   *sets[MAX_LANG];
 190     int         duplicate[MAX_LANG];
 191     int         country[MAX_LANG];
 192     char        *names[MAX_LANG];
 193     char        *langs[MAX_LANG];
 194     FILE        *f;
 195     int         ncountry = 0;
 196     int         i = 0;
 197     FcCharLeaf  **leaves, **sleaves;
 198     int         total_leaves = 0;
 199     int         l, sl, tl;
 200     int         c;
 201     char        line[1024];
 202     FcChar32    map[MAX_LANG_SET_MAP];
 203     int         num_lang_set_map;
 204
 205     while (*++argv)
 206     {
 207         if (i == MAX_LANG)
 208             fatal (*argv, 0, "Too many languages");
 209         files[i++] = *argv;
 210     }
 211     files[i] = 0;
 212     qsort (files, i, sizeof (char *), compare);
 213     i = 0;
 214     while (files[i])
 215     {
 216         f = fopen (files[i], "r");
 217         if (!f)
 218             fatal (files[i], 0, strerror (errno));
 219         sets[i] = scan (f, files[i]);
 220         names[i] = get_name (files[i]);
 221         langs[i] = get_lang(names[i]);
 222         if (strchr (langs[i], '-'))
 223             country[ncountry++] = i;
 224
 225         total_leaves += sets[i]->num;
 226         i++;
 227         fclose (f);
 228     }
 229     sets[i] = 0;
 230     leaves = malloc (total_leaves * sizeof (FcCharLeaf *));
 231     tl = 0;
 232     /*
 233      * Find unique leaves
 234      */
 235     for (i = 0; sets[i]; i++)
 236     {
 237         sleaves = sets[i]->leaves;
 238         for (sl = 0; sl < sets[i]->num; sl++)
 239         {
 240             for (l = 0; l < tl; l++)
 241                 if (leaves[l] == sleaves[sl])
 242                     break;
 243             if (l == tl)
 244                 leaves[tl++] = sleaves[sl];
 245         }
 246     }
 247
 248     /*
 249      * Scan the input until the marker is found
 250      */
 251
 252     while (fgets (line, sizeof (line), stdin))
 253     {
 254         if (!strncmp (line, "@@@", 3))
 255             break;
 256         fputs (line, stdout);
 257     }
 258
 259     printf ("/* total size: %d unique leaves: %d */\n\n",
 260             total_leaves, tl);
 261     /*
 262      * Dump leaves
 263      */
 264     printf ("static const FcCharLeaf    leaves[%d] = {\n", tl);
 265     for (l = 0; l < tl; l++)
 266     {
 267         printf ("    { { /* %d */", l);
 268         for (i = 0; i < 256/32; i++)
 269         {
 270             if (i % 4 == 0)
 271                 printf ("\n   ");
 272             printf (" 0x%08x,", leaves[l]->map[i]);
 273         }
 274         printf ("\n    } },\n");
 275     }
 276     printf ("};\n\n");
 277     printf ("#define L(n) ((FcCharLeaf *) &leaves[n])\n\n");
 278
 279     /*
 280      * Find duplicate charsets
 281      */
 282     duplicate[0] = -1;
 283     for (i = 1; sets[i]; i++)
 284     {
 285         int j;
 286
 287         duplicate[i] = -1;
 288         for (j = 0; j < i; j++)
 289             if (sets[j] == sets[i])
 290             {
 291                 duplicate[i] = j;
 292                 break;
 293             }
 294     }
 295
 296     /*
 297      * Dump arrays
 298      */
 299     for (i = 0; sets[i]; i++)
 300     {
 301         int n;
 302
 303         if (duplicate[i] >= 0)
 304             continue;
 305         printf ("static const FcCharLeaf *leaves_%s[%d] = {\n",
 306                 names[i], sets[i]->num);
 307         for (n = 0; n < sets[i]->num; n++)
 308         {
 309             if (n % 8 == 0)
 310                 printf ("   ");
 311             for (l = 0; l < tl; l++)
 312                 if (leaves[l] == sets[i]->leaves[n])
 313                     break;
 314             if (l == tl)
 315                 fatal (names[i], 0, "can't find leaf");
 316             printf (" L(%3d),", l);
 317             if (n % 8 == 7)
 318                 printf ("\n");
 319         }
 320         if (n % 8 != 0)
 321             printf ("\n");
 322         printf ("};\n\n");
 323
 324
 325         printf ("static const FcChar16 numbers_%s[%d] = {\n",
 326                 names[i], sets[i]->num);
 327         for (n = 0; n < sets[i]->num; n++)
 328         {
 329             if (n % 8 == 0)
 330                 printf ("   ");
 331             printf (" 0x%04x,", sets[i]->numbers[n]);
 332             if (n % 8 == 7)
 333                 printf ("\n");
 334         }
 335         if (n % 8 != 0)
 336             printf ("\n");
 337         printf ("};\n\n");
 338     }
 339     printf ("#undef L\n\n");
 340     /*
 341      * Dump sets
 342      */
 343     printf ("static const FcLangCharSet  fcLangCharSets[] = {\n");
 344     for (i = 0; sets[i]; i++)
 345     {
 346         int     j = duplicate[i];
 347         if (j < 0)
 348             j = i;
 349         printf ("    { (FcChar8 *) \"%s\",\n"
 350                 "      { FC_REF_CONSTANT, %d, "
 351                 "(FcCharLeaf **) leaves_%s, "
 352                 "(FcChar16 *) numbers_%s } },\n",
 353                 langs[i],
 354                 sets[j]->num, names[j], names[j]);
 355     }
 356     printf ("};\n\n");
 357     printf ("#define NUM_LANG_CHAR_SET  %d\n", i);
 358     num_lang_set_map = (i + 31) / 32;
 359     printf ("#define NUM_LANG_SET_MAP   %d\n", num_lang_set_map);
 360     /*
 361      * Dump indices with country codes
 362      */
 363     if (ncountry)
 364     {
 365         int     ncountry_ent = 0;
 366         printf ("\n");
 367         printf ("static const FcChar32 fcLangCountrySets[][NUM_LANG_SET_MAP] = {\n");
 368         for (c = 0; c < ncountry; c++)
 369         {
 370             i = country[c];
 371             if (i >= 0)
 372             {
 373                 int l = strchr (langs[i], '-') - langs[i];
 374                 int d, k;
 375
 376                 for (k = 0; k < num_lang_set_map; k++)
 377                     map[k] = 0;
 378
 379                 BitSet (map, i);
 380                 for (d = c + 1; d < ncountry; d++)
 381                 {
 382                     int j = country[d];
 383                     if (j >= 0 && !strncmp (langs[j], langs[i], l))
 384                     {
 385                         BitSet(map, j);
 386                         country[d] = -1;
 387                     }
 388                 }
 389                 printf ("    {");
 390                 for (k = 0; k < num_lang_set_map; k++)
 391                     printf (" 0x%08x,", map[k]);
 392                 printf (" }, /* %*.*s */\n",
 393                         l, l, langs[i]);
 394                 ++ncountry_ent;
 395             }
 396         }
 397         printf ("};\n\n");
 398         printf ("#define NUM_COUNTRY_SET %d\n", ncountry_ent);
 399     }
 400
 401     while (fgets (line, sizeof (line), stdin))
 402         fputs (line, stdout);
 403
 404     fflush (stdout);
 405     exit (ferror (stdout));
 406 }