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 * 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.
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 static int compare (const void *a, const void *b)
229 const FcChar8 *const *as = a, *const *bs = b;
230 return FcStrCmpIgnoreCase (*as, *bs);
233 #define MAX_LANG 1024
234 #define MAX_LANG_SET_MAP ((MAX_LANG + 31) / 32)
236 #define BitSet(map, id) ((map)[(id)>>5] |= ((FcChar32) 1 << ((id) & 0x1f)))
237 #define BitGet(map, id) ((map)[(id)>>5] >> ((id) & 0x1f)) & 1)
240 main (int argc, char **argv)
242 static char *files[MAX_LANG];
243 static const FcCharSet *sets[MAX_LANG];
244 static int duplicate[MAX_LANG];
245 static int country[MAX_LANG];
246 static char *names[MAX_LANG];
247 static char *langs[MAX_LANG];
248 static int off[MAX_LANG];
255 int total_leaves = 0;
257 static char line[1024];
258 static FcChar32 map[MAX_LANG_SET_MAP];
259 int num_lang_set_map;
260 int setRangeStart[26];
262 FcChar8 setRangeChar;
263 FcCharSetFreezer *freezer;
265 freezer = FcCharSetFreezerCreate ();
267 fatal (argv[0], 0, "out of memory");
271 if (!strcmp (argv[argi], "-d"))
278 fatal (argv[0], 0, "Too many languages");
279 files[i++] = argv[argi++];
282 qsort (files, i, sizeof (char *), compare);
286 f = scanopen (files[i]);
288 fatal (files[i], 0, strerror (errno));
289 sets[i] = scan (f, files[i], freezer);
290 names[i] = get_name (files[i]);
291 langs[i] = get_lang(names[i]);
292 if (strchr (langs[i], '-'))
293 country[ncountry++] = i;
295 total_leaves += sets[i]->num;
301 leaves = malloc (total_leaves * sizeof (FcCharLeaf *));
306 for (i = 0; sets[i]; i++)
308 for (sl = 0; sl < sets[i]->num; sl++)
310 for (l = 0; l < tl; l++)
311 if (leaves[l] == FcCharSetLeaf(sets[i], sl))
314 leaves[tl++] = FcCharSetLeaf(sets[i], sl);
319 * Scan the input until the marker is found
322 while (fgets (line, sizeof (line), stdin))
324 if (!strncmp (line, "@@@", 3))
326 fputs (line, stdout);
329 printf ("/* total size: %d unique leaves: %d */\n\n",
333 * Find duplicate charsets
336 for (i = 1; sets[i]; i++)
341 for (j = 0; j < i; j++)
342 if (sets[j] == sets[i])
350 for (i = 0; sets[i]; i++) {
351 if (duplicate[i] >= 0)
357 printf ("#define LEAF0 (%d * sizeof (FcLangCharSet))\n", nsets);
358 printf ("#define OFF0 (LEAF0 + %d * sizeof (FcCharLeaf))\n", tl);
359 printf ("#define NUM0 (OFF0 + %d * sizeof (intptr_t))\n", tn);
360 printf ("#define SET(n) (n * sizeof (FcLangCharSet) + offsetof (FcLangCharSet, charset))\n");
361 printf ("#define OFF(s,o) (OFF0 + o * sizeof (intptr_t) - SET(s))\n");
362 printf ("#define NUM(s,n) (NUM0 + n * sizeof (FcChar16) - SET(s))\n");
363 printf ("#define LEAF(o,l) (LEAF0 + l * sizeof (FcCharLeaf) - (OFF0 + o * sizeof (intptr_t)))\n");
364 printf ("#define fcLangCharSets (fcLangData.langCharSets)\n");
367 printf ("static const struct {\n"
368 " FcLangCharSet langCharSets[%d];\n"
369 " FcCharLeaf leaves[%d];\n"
370 " intptr_t leaf_offsets[%d];\n"
371 " FcChar16 numbers[%d];\n"
372 "} fcLangData = {\n",
380 for (i = 0; sets[i]; i++)
382 int j = duplicate[i];
387 printf (" { \"%s\", "
388 " { FC_REF_CONSTANT, %d, OFF(%d,%d), NUM(%d,%d) } }, /* %d */\n",
390 sets[j]->num, i, off[j], i, off[j], i);
398 for (l = 0; l < tl; l++)
400 printf (" { { /* %d */", l);
401 for (i = 0; i < 256/32; i++)
405 printf (" 0x%08x,", leaves[l]->map[i]);
407 printf ("\n } },\n");
415 for (i = 0; sets[i]; i++)
419 if (duplicate[i] >= 0)
421 printf (" /* %s */\n", names[i]);
422 for (n = 0; n < sets[i]->num; n++)
426 for (l = 0; l < tl; l++)
427 if (leaves[l] == FcCharSetLeaf(sets[i], n))
430 fatal (names[i], 0, "can't find leaf");
431 printf (" LEAF(%3d,%3d),", off[i], l);
442 for (i = 0; sets[i]; i++)
446 if (duplicate[i] >= 0)
448 printf (" /* %s */\n", names[i]);
449 for (n = 0; n < sets[i]->num; n++)
453 printf (" 0x%04x,", FcCharSetNumbers (sets[i])[n]);
464 printf ("#define NUM_LANG_CHAR_SET %d\n", i);
465 num_lang_set_map = (i + 31) / 32;
466 printf ("#define NUM_LANG_SET_MAP %d\n", num_lang_set_map);
468 * Dump indices with country codes
473 int ncountry_ent = 0;
475 printf ("static const FcChar32 fcLangCountrySets[][NUM_LANG_SET_MAP] = {\n");
476 for (c = 0; c < ncountry; c++)
481 int lang = strchr (langs[i], '-') - langs[i];
484 for (k = 0; k < num_lang_set_map; k++)
488 for (d = c + 1; d < ncountry; d++)
491 if (j >= 0 && !strncmp (langs[j], langs[i], l))
498 for (k = 0; k < num_lang_set_map; k++)
499 printf (" 0x%08x,", map[k]);
500 printf (" }, /* %*.*s */\n",
501 lang, lang, langs[i]);
506 printf ("#define NUM_COUNTRY_SET %d\n", ncountry_ent);
511 * Find ranges for each letter for faster searching
514 memset(setRangeStart, '\0', sizeof (setRangeStart));
515 memset(setRangeEnd, '\0', sizeof (setRangeEnd));
516 for (i = 0; sets[i]; i++)
518 char c = names[i][0];
520 while (setRangeChar <= c && c <= 'z')
521 setRangeStart[setRangeChar++ - 'a'] = i;
523 for (setRangeChar = 'a'; setRangeChar < 'z'; setRangeChar++)
524 setRangeEnd[setRangeChar - 'a'] = setRangeStart[setRangeChar+1-'a'] - 1;
525 setRangeEnd[setRangeChar - 'a'] = i - 1;
528 * Dump sets start/finish for the fastpath
530 printf ("static const FcLangCharSetRange fcLangCharSetRanges[] = {\n");
531 for (setRangeChar = 'a'; setRangeChar <= 'z' ; setRangeChar++)
533 printf (" { %d, %d }, /* %c */\n",
534 setRangeStart[setRangeChar - 'a'],
535 setRangeEnd[setRangeChar - 'a'], setRangeChar);
539 while (fgets (line, sizeof (line), stdin))
540 fputs (line, stdout);
543 exit (ferror (stdout));