2 * $RCSId: xc/lib/fontconfig/fc-lang/fc-lang.c,v 1.3 2002/08/22 07:36:43 keithp Exp $
4 * Copyright © 2002 Keith Packard
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.
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.
26 #include "fccharset.c"
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
41 FcMemAlloc (int kind, int size)
46 FcMemFree (int kind, int size)
51 FcCacheBankToIndex (int bank)
59 return (FcChar8 *) getenv ("HOME");
63 fatal (const char *file, int lineno, const char *msg)
66 fprintf (stderr, "%s:%d: %s\n", file, lineno, msg);
68 fprintf (stderr, "%s:%d: %s\n", file, lineno, msg);
73 get_line (FILE *f, char *line, int *lineno)
76 if (!fgets (line, 1024, f))
79 hash = strchr (line, '#');
82 if (line[0] == '\0' || line[0] == '\n' || line[0] == '\032' || line[0] == '\r')
83 return get_line (f, line, lineno);
94 f = fopen (file, "r");
102 f = fopen (path, "r");
108 * build a single charset from a source file
110 * The file format is quite simple, either
111 * a single hex value or a pair separated with a dash
113 * Comments begin with '#'
117 scan (FILE *f, char *file)
121 int start, end, ucs4;
125 while (get_line (f, line, &lineno))
127 if (!strncmp (line, "include", 7))
129 file = strchr (line, ' ');
133 if (file[end-1] == '\n')
137 fatal (file, 0, "can't open");
142 if (strchr (line, '-'))
144 if (sscanf (line, "%x-%x", &start, &end) != 2)
145 fatal (file, lineno, "parse error");
149 if (sscanf (line, "%x", &start) != 1)
150 fatal (file, lineno, "parse error");
154 c = FcCharSetCreate ();
155 for (ucs4 = start; ucs4 <= end; ucs4++)
157 if (!FcCharSetAddChar (c, ucs4))
158 fatal (file, lineno, "out of memory");
161 n = FcCharSetFreeze (c);
162 FcCharSetDestroy (c);
167 * Convert a file name into a name suitable for C declarations
170 get_name (char *file)
175 dot = strchr (file, '.');
177 dot = file + strlen(file);
178 name = malloc (dot - file + 1);
179 strncpy (name, file, dot - file);
180 name[dot-file] = '\0';
185 * Convert a C name into a language name
188 get_lang (char *name)
190 char *lang = malloc (strlen (name) + 1);
194 while ((c = *name++))
196 if (isupper ((int) (unsigned char) c))
197 c = tolower ((int) (unsigned char) c);
208 static int compare (const void *a, const void *b)
210 const FcChar8 *const *as = a, *const *bs = b;
211 return FcStrCmpIgnoreCase (*as, *bs);
214 #define MAX_LANG 1024
215 #define MAX_LANG_SET_MAP ((MAX_LANG + 31) / 32)
217 #define BitSet(map, id) ((map)[(id)>>5] |= ((FcChar32) 1 << ((id) & 0x1f)))
218 #define BitGet(map, id) ((map)[(id)>>5] >> ((id) & 0x1f)) & 1)
221 main (int argc, char **argv)
223 char *files[MAX_LANG];
224 FcCharSet *sets[MAX_LANG];
225 int duplicate[MAX_LANG];
226 int country[MAX_LANG];
227 char *names[MAX_LANG];
228 char *langs[MAX_LANG];
234 int total_leaves = 0;
238 FcChar32 map[MAX_LANG_SET_MAP];
239 int num_lang_set_map;
240 int setRangeStart[26];
242 FcChar8 setRangeChar;
247 if (!strcmp (argv[argi], "-d"))
254 fatal (argv[0], 0, "Too many languages");
255 files[i++] = argv[argi++];
258 qsort (files, i, sizeof (char *), compare);
262 f = scanopen (files[i]);
264 fatal (files[i], 0, strerror (errno));
265 sets[i] = scan (f, files[i]);
266 names[i] = get_name (files[i]);
267 langs[i] = get_lang(names[i]);
268 if (strchr (langs[i], '-'))
269 country[ncountry++] = i;
271 total_leaves += sets[i]->num;
276 leaves = malloc (total_leaves * sizeof (FcCharLeaf *));
281 for (i = 0; sets[i]; i++)
283 for (sl = 0; sl < sets[i]->num; sl++)
285 for (l = 0; l < tl; l++)
286 if (leaves[l] == FcCharSetGetLeaf(sets[i], sl))
289 leaves[tl++] = FcCharSetGetLeaf(sets[i], sl);
294 * Scan the input until the marker is found
297 while (fgets (line, sizeof (line), stdin))
299 if (!strncmp (line, "@@@", 3))
301 fputs (line, stdout);
304 printf ("/* total size: %d unique leaves: %d */\n\n",
309 printf ("static const FcCharLeaf leaves[%d] = {\n", tl);
310 for (l = 0; l < tl; l++)
312 printf (" { { /* %d */", l);
313 for (i = 0; i < 256/32; i++)
317 printf (" 0x%08x,", leaves[l]->map[i]);
319 printf ("\n } },\n");
322 printf ("#define L(n) ((FcCharLeaf *) &leaves[n])\n\n");
325 * Find duplicate charsets
328 for (i = 1; sets[i]; i++)
333 for (j = 0; j < i; j++)
334 if (sets[j] == sets[i])
342 * Find ranges for each letter for faster searching
345 for (i = 0; sets[i]; i++)
347 char c = names[i][0];
349 while (setRangeChar <= c && c <= 'z')
350 setRangeStart[setRangeChar++ - 'a'] = i;
352 for (setRangeChar = 'a'; setRangeChar < 'z'; setRangeChar++)
353 setRangeEnd[setRangeChar - 'a'] = setRangeStart[setRangeChar+1-'a'] - 1;
354 setRangeEnd[setRangeChar - 'a'] = i - 1;
359 for (i = 0; sets[i]; i++)
363 if (duplicate[i] >= 0)
365 printf ("static const FcCharLeaf *leaves_%s[%d] = {\n",
366 names[i], sets[i]->num);
367 for (n = 0; n < sets[i]->num; n++)
371 for (l = 0; l < tl; l++)
372 if (leaves[l] == FcCharSetGetLeaf(sets[i], n))
375 fatal (names[i], 0, "can't find leaf");
376 printf (" L(%3d),", l);
385 printf ("static const FcChar16 numbers_%s[%d] = {\n",
386 names[i], sets[i]->num);
387 for (n = 0; n < sets[i]->num; n++)
391 printf (" 0x%04x,", FcCharSetGetNumbers(sets[i])[n]);
399 printf ("#undef L\n\n");
405 printf ("static const FcLangCharSet fcLangCharSets[] = {\n");
406 for (i = 0; sets[i]; i++)
408 int j = duplicate[i];
412 printf (" { (FcChar8 *) \"%s\",\n"
413 " { FC_REF_CONSTANT, %d, FC_BANK_DYNAMIC, "
414 "{ { (FcCharLeaf **) leaves_%s, "
415 "(FcChar16 *) numbers_%s } } } },\n",
417 sets[j]->num, names[j], names[j]);
420 printf ("#define NUM_LANG_CHAR_SET %d\n", i);
421 num_lang_set_map = (i + 31) / 32;
422 printf ("#define NUM_LANG_SET_MAP %d\n", num_lang_set_map);
424 * Dump indices with country codes
428 int ncountry_ent = 0;
430 printf ("static const FcChar32 fcLangCountrySets[][NUM_LANG_SET_MAP] = {\n");
431 for (c = 0; c < ncountry; c++)
436 int l = strchr (langs[i], '-') - langs[i];
439 for (k = 0; k < num_lang_set_map; k++)
443 for (d = c + 1; d < ncountry; d++)
446 if (j >= 0 && !strncmp (langs[j], langs[i], l))
453 for (k = 0; k < num_lang_set_map; k++)
454 printf (" 0x%08x,", map[k]);
455 printf (" }, /* %*.*s */\n",
461 printf ("#define NUM_COUNTRY_SET %d\n", ncountry_ent);
466 * Dump sets start/finish for the fastpath
468 printf ("static const FcLangCharSetRange fcLangCharSetRanges[] = {\n");
469 for (setRangeChar = 'a'; setRangeChar <= 'z' ; setRangeChar++)
471 printf (" { %d, %d }, /* %c */\n",
472 setRangeStart[setRangeChar - 'a'],
473 setRangeEnd[setRangeChar - 'a'], setRangeChar);
477 while (fgets (line, sizeof (line), stdin))
478 fputs (line, stdout);
481 exit (ferror (stdout));