2 * fontconfig/fc-lang/fc-lang.c
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 * THE AUTHOR(S) DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
17 * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
18 * EVENT SHALL THE AUTHOR(S) 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.
25 #include "fccharset.c"
27 #include "fcserialize.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 FcCacheObjectReference (void *object)
56 FcCacheObjectDereference (void *object)
65 return (FcChar8 *) getenv ("HOME");
69 fatal (const char *file, int lineno, const char *msg)
72 fprintf (stderr, "%s:%d: %s\n", file, lineno, msg);
74 fprintf (stderr, "%s: %s\n", file, msg);
79 get_line (FILE *f, char *line, int *lineno)
83 if (!fgets (line, 1024, f))
86 hash = strchr (line, '#');
91 while (end > 0 && isspace (line[end-1]))
94 if (line[0] == '\0' || line[0] == '\n' || line[0] == '\032' || line[0] == '\r')
95 return get_line (f, line, lineno);
102 scanopen (char *file)
106 f = fopen (file, "r");
114 f = fopen (path, "r");
120 * build a single charset from a source file
122 * The file format is quite simple, either
123 * a single hex value or a pair separated with a dash
125 * Comments begin with '#'
128 static const FcCharSet *
129 scan (FILE *f, char *file, FcCharSetFreezer *freezer)
133 int start, end, ucs4;
137 while (get_line (f, line, &lineno))
139 if (!strncmp (line, "include", 7))
143 included_file = strchr (line, ' ');
146 "invalid syntax, expected: include filename");
147 while (isspace(*included_file))
149 included_f = scanopen (included_file);
151 fatal (included_file, 0, "can't open");
152 n = scan (included_f, included_file, freezer);
155 c = FcCharSetCreate ();
156 if (!FcCharSetMerge (c, n, NULL))
157 fatal (file, lineno, "out of memory");
158 FcCharSetDestroy (n);
161 if (strchr (line, '-'))
163 if (sscanf (line, "%x-%x", &start, &end) != 2)
164 fatal (file, lineno, "parse error");
168 if (sscanf (line, "%x", &start) != 1)
169 fatal (file, lineno, "parse error");
173 c = FcCharSetCreate ();
174 for (ucs4 = start; ucs4 <= end; ucs4++)
176 if (!FcCharSetAddChar (c, ucs4))
177 fatal (file, lineno, "out of memory");
180 n = FcCharSetFreeze (freezer, c);
181 FcCharSetDestroy (c);
186 * Convert a file name into a name suitable for C declarations
189 get_name (char *file)
194 dot = strchr (file, '.');
196 dot = file + strlen(file);
197 name = malloc (dot - file + 1);
198 strncpy (name, file, dot - file);
199 name[dot-file] = '\0';
204 * Convert a C name into a language name
207 get_lang (char *name)
209 char *lang = malloc (strlen (name) + 1);
213 while ((c = *name++))
215 if (isupper ((int) (unsigned char) c))
216 c = tolower ((int) (unsigned char) c);
227 typedef struct _Entry {
232 static int compare (const void *a, const void *b)
234 const Entry const *as = a, *bs = b;
235 return FcStrCmpIgnoreCase (as->file, bs->file);
238 #define MAX_LANG 1024
239 #define MAX_LANG_SET_MAP ((MAX_LANG + 31) / 32)
241 #define BitSet(map, i) ((map)[(entries[i].id)>>5] |= ((FcChar32) 1 << ((entries[i].id) & 0x1f)))
242 #define BitGet(map, i) ((map)[(entries[i].id)>>5] >> ((entries[i].id) & 0x1f)) & 1)
245 main (int argc, char **argv)
247 static Entry entries[MAX_LANG];
248 static const FcCharSet *sets[MAX_LANG];
249 static int duplicate[MAX_LANG];
250 static int country[MAX_LANG];
251 static char *names[MAX_LANG];
252 static char *langs[MAX_LANG];
253 static int off[MAX_LANG];
260 int total_leaves = 0;
262 static char line[1024];
263 static FcChar32 map[MAX_LANG_SET_MAP];
264 int num_lang_set_map;
265 int setRangeStart[26];
267 FcChar8 setRangeChar;
268 FcCharSetFreezer *freezer;
270 freezer = FcCharSetFreezerCreate ();
272 fatal (argv[0], 0, "out of memory");
276 if (!strcmp (argv[argi], "-d"))
283 fatal (argv[0], 0, "Too many languages");
285 entries[i].file = argv[argi++];
289 qsort (entries, i, sizeof (Entry), compare);
291 while (entries[i].file)
293 f = scanopen (entries[i].file);
295 fatal (entries[i].file, 0, strerror (errno));
296 sets[i] = scan (f, entries[i].file, freezer);
297 names[i] = get_name (entries[i].file);
298 langs[i] = get_lang(names[i]);
299 if (strchr (langs[i], '-'))
300 country[ncountry++] = i;
302 total_leaves += sets[i]->num;
308 leaves = malloc (total_leaves * sizeof (FcCharLeaf *));
313 for (i = 0; sets[i]; i++)
315 for (sl = 0; sl < sets[i]->num; sl++)
317 for (l = 0; l < tl; l++)
318 if (leaves[l] == FcCharSetLeaf(sets[i], sl))
321 leaves[tl++] = FcCharSetLeaf(sets[i], sl);
326 * Scan the input until the marker is found
329 while (fgets (line, sizeof (line), stdin))
331 if (!strncmp (line, "@@@", 3))
333 fputs (line, stdout);
336 printf ("/* total size: %d unique leaves: %d */\n\n",
340 * Find duplicate charsets
343 for (i = 1; sets[i]; i++)
348 for (j = 0; j < i; j++)
349 if (sets[j] == sets[i])
357 for (i = 0; sets[i]; i++) {
358 if (duplicate[i] >= 0)
364 printf ("#define LEAF0 (%d * sizeof (FcLangCharSet))\n", nsets);
365 printf ("#define OFF0 (LEAF0 + %d * sizeof (FcCharLeaf))\n", tl);
366 printf ("#define NUM0 (OFF0 + %d * sizeof (uintptr_t))\n", tn);
367 printf ("#define SET(n) (n * sizeof (FcLangCharSet) + offsetof (FcLangCharSet, charset))\n");
368 printf ("#define OFF(s,o) (OFF0 + o * sizeof (uintptr_t) - SET(s))\n");
369 printf ("#define NUM(s,n) (NUM0 + n * sizeof (FcChar16) - SET(s))\n");
370 printf ("#define LEAF(o,l) (LEAF0 + l * sizeof (FcCharLeaf) - (OFF0 + o * sizeof (intptr_t)))\n");
371 printf ("#define fcLangCharSets (fcLangData.langCharSets)\n");
372 printf ("#define fcLangCharSetIndices (fcLangData.langIndices)\n");
373 printf ("#define fcLangCharSetIndicesInv (fcLangData.langIndicesInv)\n");
376 printf ("static const struct {\n"
377 " FcLangCharSet langCharSets[%d];\n"
378 " FcCharLeaf leaves[%d];\n"
379 " uintptr_t leaf_offsets[%d];\n"
380 " FcChar16 numbers[%d];\n"
381 " FcChar%s langIndices[%d];\n"
382 " FcChar%s langIndicesInv[%d];\n"
383 "} fcLangData = {\n",
385 nsets < 256 ? "8 " : "16", nsets, nsets < 256 ? "8 " : "16", nsets);
392 for (i = 0; sets[i]; i++)
394 int j = duplicate[i];
399 printf (" { \"%s\", "
400 " { FC_REF_CONSTANT, %d, OFF(%d,%d), NUM(%d,%d) } }, /* %d */\n",
402 sets[j]->num, i, off[j], i, off[j], i);
410 for (l = 0; l < tl; l++)
412 printf (" { { /* %d */", l);
413 for (i = 0; i < 256/32; i++)
417 printf (" 0x%08x,", leaves[l]->map[i]);
419 printf ("\n } },\n");
427 for (i = 0; sets[i]; i++)
431 if (duplicate[i] >= 0)
433 printf (" /* %s */\n", names[i]);
434 for (n = 0; n < sets[i]->num; n++)
438 for (l = 0; l < tl; l++)
439 if (leaves[l] == FcCharSetLeaf(sets[i], n))
442 fatal (names[i], 0, "can't find leaf");
443 printf (" LEAF(%3d,%3d),", off[i], l);
454 for (i = 0; sets[i]; i++)
458 if (duplicate[i] >= 0)
460 printf (" /* %s */\n", names[i]);
461 for (n = 0; n < sets[i]->num; n++)
465 printf (" 0x%04x,", FcCharSetNumbers (sets[i])[n]);
476 for (i = 0; sets[i]; i++)
478 printf (" %d, /* %s */\n", entries[i].id, names[i]);
485 static int entries_inv[MAX_LANG];
486 for (i = 0; sets[i]; i++)
487 entries_inv[entries[i].id] = i;
488 for (i = 0; sets[i]; i++)
489 printf (" %d, /* %s */\n", entries_inv[i], names[entries_inv[i]]);
495 printf ("#define NUM_LANG_CHAR_SET %d\n", i);
496 num_lang_set_map = (i + 31) / 32;
497 printf ("#define NUM_LANG_SET_MAP %d\n", num_lang_set_map);
499 * Dump indices with country codes
504 int ncountry_ent = 0;
506 printf ("static const FcChar32 fcLangCountrySets[][NUM_LANG_SET_MAP] = {\n");
507 for (c = 0; c < ncountry; c++)
512 int lang = strchr (langs[i], '-') - langs[i];
515 for (k = 0; k < num_lang_set_map; k++)
519 for (d = c + 1; d < ncountry; d++)
522 if (j >= 0 && !strncmp (langs[j], langs[i], lang + 1))
529 for (k = 0; k < num_lang_set_map; k++)
530 printf (" 0x%08x,", map[k]);
531 printf (" }, /* %*.*s */\n",
532 lang, lang, langs[i]);
537 printf ("#define NUM_COUNTRY_SET %d\n", ncountry_ent);
542 * Find ranges for each letter for faster searching
545 memset(setRangeStart, '\0', sizeof (setRangeStart));
546 memset(setRangeEnd, '\0', sizeof (setRangeEnd));
547 for (i = 0; sets[i]; i++)
549 char c = names[i][0];
551 while (setRangeChar <= c && c <= 'z')
552 setRangeStart[setRangeChar++ - 'a'] = i;
554 for (setRangeChar = 'a'; setRangeChar < 'z'; setRangeChar++)
555 setRangeEnd[setRangeChar - 'a'] = setRangeStart[setRangeChar+1-'a'] - 1;
556 setRangeEnd[setRangeChar - 'a'] = i - 1;
559 * Dump sets start/finish for the fastpath
562 printf ("static const FcLangCharSetRange fcLangCharSetRanges[] = {\n");
564 for (setRangeChar = 'a'; setRangeChar <= 'z' ; setRangeChar++)
566 printf (" { %d, %d }, /* %c */\n",
567 setRangeStart[setRangeChar - 'a'],
568 setRangeEnd[setRangeChar - 'a'], setRangeChar);
572 while (fgets (line, sizeof (line), stdin))
573 fputs (line, stdout);
576 exit (ferror (stdout));