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)
53 return getenv ("HOME");
57 fatal (char *file, int lineno, char *msg)
59 fprintf (stderr, "%s:%d: %s\n", file, lineno, msg);
64 get_line (FILE *f, char *line, int *lineno)
67 if (!fgets (line, 1024, f))
70 hash = strchr (line, '#');
73 if (line[0] == '\0' || line[0] == '\n' || line[0] == '\032' || line[0] == '\r')
74 return get_line (f, line, lineno);
79 * build a single charset from a source file
81 * The file format is quite simple, either
82 * a single hex value or a pair separated with a dash
84 * Comments begin with '#'
88 scan (FILE *f, char *file)
96 while (get_line (f, line, &lineno))
98 if (!strncmp (line, "include", 7))
100 file = strchr (line, ' ');
104 if (file[end-1] == '\n')
106 f = fopen (file, "r");
108 fatal (file, 0, "can't open");
113 if (strchr (line, '-'))
115 if (sscanf (line, "%x-%x", &start, &end) != 2)
116 fatal (file, lineno, "parse error");
120 if (sscanf (line, "%x", &start) != 1)
121 fatal (file, lineno, "parse error");
125 c = FcCharSetCreate ();
126 for (ucs4 = start; ucs4 <= end; ucs4++)
128 if (!FcCharSetAddChar (c, ucs4))
129 fatal (file, lineno, "out of memory");
132 n = FcCharSetFreeze (c);
133 FcCharSetDestroy (c);
138 * Convert a file name into a name suitable for C declarations
141 get_name (char *file)
146 dot = strchr (file, '.');
148 dot = file + strlen(file);
149 name = malloc (dot - file + 1);
150 strncpy (name, file, dot - file);
151 name[dot-file] = '\0';
156 * Convert a C name into a language name
159 get_lang (char *name)
161 char *lang = malloc (strlen (name) + 1);
165 while ((c = *name++))
167 if (isupper ((int) (unsigned char) c))
168 c = tolower ((int) (unsigned char) c);
179 static int compare (const void *a, const void *b)
181 const FcChar8 *const *as = a, *const *bs = b;
182 return FcStrCmpIgnoreCase (*as, *bs);
185 #define MAX_LANG 1024
186 #define MAX_LANG_SET_MAP ((MAX_LANG + 31) / 32)
188 #define BitSet(map, id) ((map)[(id)>>5] |= ((FcChar32) 1 << ((id) & 0x1f)))
189 #define BitGet(map, id) ((map)[(id)>>5] >> ((id) & 0x1f)) & 1)
192 main (int argc, char **argv)
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];
203 FcCharLeaf **leaves, **sleaves;
204 int total_leaves = 0;
208 FcChar32 map[MAX_LANG_SET_MAP];
209 int num_lang_set_map;
210 int setRangeStart[26];
212 FcChar8 setRangeChar;
217 fatal (*argv, 0, "Too many languages");
221 qsort (files, i, sizeof (char *), compare);
225 f = fopen (files[i], "r");
227 fatal (files[i], 0, strerror (errno));
228 sets[i] = scan (f, files[i]);
229 names[i] = get_name (files[i]);
230 langs[i] = get_lang(names[i]);
231 if (strchr (langs[i], '-'))
232 country[ncountry++] = i;
234 total_leaves += sets[i]->num;
239 leaves = malloc (total_leaves * sizeof (FcCharLeaf *));
244 for (i = 0; sets[i]; i++)
246 sleaves = sets[i]->leaves;
247 for (sl = 0; sl < sets[i]->num; sl++)
249 for (l = 0; l < tl; l++)
250 if (leaves[l] == sleaves[sl])
253 leaves[tl++] = sleaves[sl];
258 * Scan the input until the marker is found
261 while (fgets (line, sizeof (line), stdin))
263 if (!strncmp (line, "@@@", 3))
265 fputs (line, stdout);
268 printf ("/* total size: %d unique leaves: %d */\n\n",
273 printf ("static const FcCharLeaf leaves[%d] = {\n", tl);
274 for (l = 0; l < tl; l++)
276 printf (" { { /* %d */", l);
277 for (i = 0; i < 256/32; i++)
281 printf (" 0x%08x,", leaves[l]->map[i]);
283 printf ("\n } },\n");
286 printf ("#define L(n) ((FcCharLeaf *) &leaves[n])\n\n");
289 * Find duplicate charsets
292 for (i = 1; sets[i]; i++)
297 for (j = 0; j < i; j++)
298 if (sets[j] == sets[i])
306 * Find ranges for each letter for faster searching
309 for (i = 0; sets[i]; i++)
311 char c = names[i][0];
313 while (setRangeChar <= c && c <= 'z')
314 setRangeStart[setRangeChar++ - 'a'] = i;
316 for (setRangeChar = 'a'; setRangeChar < 'z'; setRangeChar++)
317 setRangeEnd[setRangeChar - 'a'] = setRangeStart[setRangeChar+1-'a'] - 1;
318 setRangeEnd[setRangeChar - 'a'] = i - 1;
323 for (i = 0; sets[i]; i++)
327 if (duplicate[i] >= 0)
329 printf ("static const FcCharLeaf *leaves_%s[%d] = {\n",
330 names[i], sets[i]->num);
331 for (n = 0; n < sets[i]->num; n++)
335 for (l = 0; l < tl; l++)
336 if (leaves[l] == sets[i]->leaves[n])
339 fatal (names[i], 0, "can't find leaf");
340 printf (" L(%3d),", l);
349 printf ("static const FcChar16 numbers_%s[%d] = {\n",
350 names[i], sets[i]->num);
351 for (n = 0; n < sets[i]->num; n++)
355 printf (" 0x%04x,", sets[i]->numbers[n]);
363 printf ("#undef L\n\n");
369 printf ("static const FcLangCharSet fcLangCharSets[] = {\n");
370 for (i = 0; sets[i]; i++)
372 int j = duplicate[i];
376 printf (" { (FcChar8 *) \"%s\",\n"
377 " { FC_REF_CONSTANT, %d, "
378 "(FcCharLeaf **) leaves_%s, "
379 "(FcChar16 *) numbers_%s } },\n",
381 sets[j]->num, names[j], names[j]);
384 printf ("#define NUM_LANG_CHAR_SET %d\n", i);
385 num_lang_set_map = (i + 31) / 32;
386 printf ("#define NUM_LANG_SET_MAP %d\n", num_lang_set_map);
388 * Dump indices with country codes
392 int ncountry_ent = 0;
394 printf ("static const FcChar32 fcLangCountrySets[][NUM_LANG_SET_MAP] = {\n");
395 for (c = 0; c < ncountry; c++)
400 int l = strchr (langs[i], '-') - langs[i];
403 for (k = 0; k < num_lang_set_map; k++)
407 for (d = c + 1; d < ncountry; d++)
410 if (j >= 0 && !strncmp (langs[j], langs[i], l))
417 for (k = 0; k < num_lang_set_map; k++)
418 printf (" 0x%08x,", map[k]);
419 printf (" }, /* %*.*s */\n",
425 printf ("#define NUM_COUNTRY_SET %d\n", ncountry_ent);
430 * Dump sets start/finish for the fastpath
432 printf ("static const FcLangCharSetRange fcLangCharSetRanges[] = {\n");
433 for (setRangeChar = 'a'; setRangeChar <= 'z' ; setRangeChar++)
435 printf (" { %d, %d }, /* %c */\n",
436 setRangeStart[setRangeChar - 'a'],
437 setRangeEnd[setRangeChar - 'a'], setRangeChar);
441 while (fgets (line, sizeof (line), stdin))
442 fputs (line, stdout);
445 exit (ferror (stdout));