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
40 const FcChar16 langBankNumbers
[1]; /* place holders so that externs resolve */
41 const FcCharLeaf langBankLeaves
[1];
42 const int langBankLeafIdx
[1];
45 FcMemAlloc (int kind
, int size
)
50 FcMemFree (int kind
, int size
)
56 FcValueList
** _fcValueLists
= 0;
57 FcPatternElt
** _fcPatternElts
= 0;
61 FcCacheBankToIndexMTF (int bank
)
69 return (FcChar8
*) getenv ("HOME");
73 fatal (const char *file
, int lineno
, const char *msg
)
76 fprintf (stderr
, "%s:%d: %s\n", file
, lineno
, msg
);
78 fprintf (stderr
, "%s: %s\n", file
, msg
);
83 get_line (FILE *f
, char *line
, int *lineno
)
87 if (!fgets (line
, 1024, f
))
90 hash
= strchr (line
, '#');
95 while (end
> 0 && isspace (line
[end
-1]))
98 if (line
[0] == '\0' || line
[0] == '\n' || line
[0] == '\032' || line
[0] == '\r')
99 return get_line (f
, line
, lineno
);
103 static char *dir
= 0;
106 scanopen (char *file
)
110 f
= fopen (file
, "r");
118 f
= fopen (path
, "r");
124 * build a single charset from a source file
126 * The file format is quite simple, either
127 * a single hex value or a pair separated with a dash
129 * Comments begin with '#'
133 scan (FILE *f
, char *file
)
137 int start
, end
, ucs4
;
141 while (get_line (f
, line
, &lineno
))
143 if (!strncmp (line
, "include", 7))
145 file
= strchr (line
, ' ');
146 while (isspace(*file
))
150 fatal (file
, 0, "can't open");
155 if (strchr (line
, '-'))
157 if (sscanf (line
, "%x-%x", &start
, &end
) != 2)
158 fatal (file
, lineno
, "parse error");
162 if (sscanf (line
, "%x", &start
) != 1)
163 fatal (file
, lineno
, "parse error");
167 c
= FcCharSetCreate ();
168 for (ucs4
= start
; ucs4
<= end
; ucs4
++)
170 if (!FcCharSetAddChar (c
, ucs4
))
171 fatal (file
, lineno
, "out of memory");
174 n
= FcCharSetFreeze (c
);
175 FcCharSetDestroy (c
);
180 * Convert a file name into a name suitable for C declarations
183 get_name (char *file
)
188 dot
= strchr (file
, '.');
190 dot
= file
+ strlen(file
);
191 name
= malloc (dot
- file
+ 1);
192 strncpy (name
, file
, dot
- file
);
193 name
[dot
-file
] = '\0';
198 * Convert a C name into a language name
201 get_lang (char *name
)
203 char *lang
= malloc (strlen (name
) + 1);
207 while ((c
= *name
++))
209 if (isupper ((int) (unsigned char) c
))
210 c
= tolower ((int) (unsigned char) c
);
221 static int compare (const void *a
, const void *b
)
223 const FcChar8
*const *as
= a
, *const *bs
= b
;
224 return FcStrCmpIgnoreCase (*as
, *bs
);
227 #define MAX_LANG 1024
228 #define MAX_LANG_SET_MAP ((MAX_LANG + 31) / 32)
230 #define BitSet(map, id) ((map)[(id)>>5] |= ((FcChar32) 1 << ((id) & 0x1f)))
231 #define BitGet(map, id) ((map)[(id)>>5] >> ((id) & 0x1f)) & 1)
234 main (int argc
, char **argv
)
236 static char *files
[MAX_LANG
];
237 static FcCharSet
*sets
[MAX_LANG
];
238 static int duplicate
[MAX_LANG
];
239 static int offsets
[MAX_LANG
];
240 static int country
[MAX_LANG
];
241 static char *names
[MAX_LANG
];
242 static char *langs
[MAX_LANG
];
249 int total_leaves
= 0;
250 int offset_count
= 0;
252 static char line
[1024];
253 static FcChar32 map
[MAX_LANG_SET_MAP
];
254 int num_lang_set_map
;
255 int setRangeStart
[26];
257 FcChar8 setRangeChar
;
262 if (!strcmp (argv
[argi
], "-d"))
269 fatal (argv
[0], 0, "Too many languages");
270 files
[i
++] = argv
[argi
++];
273 qsort (files
, i
, sizeof (char *), compare
);
277 f
= scanopen (files
[i
]);
279 fatal (files
[i
], 0, strerror (errno
));
280 sets
[i
] = scan (f
, files
[i
]);
281 names
[i
] = get_name (files
[i
]);
282 langs
[i
] = get_lang(names
[i
]);
283 if (strchr (langs
[i
], '-'))
284 country
[ncountry
++] = i
;
286 total_leaves
+= sets
[i
]->num
;
291 leaves
= malloc (total_leaves
* sizeof (FcCharLeaf
*));
296 for (i
= 0; sets
[i
]; i
++)
298 for (sl
= 0; sl
< sets
[i
]->num
; sl
++)
300 for (l
= 0; l
< tl
; l
++)
301 if (leaves
[l
] == FcCharSetGetLeaf(sets
[i
], sl
))
304 leaves
[tl
++] = FcCharSetGetLeaf(sets
[i
], sl
);
309 * Scan the input until the marker is found
312 while (fgets (line
, sizeof (line
), stdin
))
314 if (!strncmp (line
, "@@@", 3))
316 fputs (line
, stdout
);
319 printf ("/* total size: %d unique leaves: %d */\n\n",
324 printf ("const FcCharLeaf langBankLeaves[%d] = {\n", tl
);
325 for (l
= 0; l
< tl
; l
++)
327 printf (" { { /* %d */", l
);
328 for (i
= 0; i
< 256/32; i
++)
332 printf (" 0x%08x,", leaves
[l
]->map
[i
]);
334 printf ("\n } },\n");
339 * Find duplicate charsets
342 for (i
= 1; sets
[i
]; i
++)
347 for (j
= 0; j
< i
; j
++)
348 if (sets
[j
] == sets
[i
])
356 * Find ranges for each letter for faster searching
359 for (i
= 0; sets
[i
]; i
++)
361 char c
= names
[i
][0];
363 while (setRangeChar
<= c
&& c
<= 'z')
364 setRangeStart
[setRangeChar
++ - 'a'] = i
;
366 for (setRangeChar
= 'a'; setRangeChar
< 'z'; setRangeChar
++)
367 setRangeEnd
[setRangeChar
- 'a'] = setRangeStart
[setRangeChar
+1-'a'] - 1;
368 setRangeEnd
[setRangeChar
- 'a'] = i
- 1;
373 for (i
= 0; sets
[i
]; i
++)
377 if (duplicate
[i
] >= 0)
380 for (n
= 0; n
< sets
[i
]->num
; n
++)
382 for (l
= 0; l
< tl
; l
++)
383 if (leaves
[l
] == FcCharSetGetLeaf(sets
[i
], n
))
386 fatal (names
[i
], 0, "can't find leaf");
390 offset
+= sets
[i
]->num
;
393 printf ("const int langBankLeafIdx[%d] = {\n",
395 for (i
= 0; sets
[i
]; i
++)
399 if (duplicate
[i
] >= 0)
401 for (n
= 0; n
< sets
[i
]->num
; n
++)
405 for (l
= 0; l
< tl
; l
++)
406 if (leaves
[l
] == FcCharSetGetLeaf(sets
[i
], n
))
409 fatal (names
[i
], 0, "can't find leaf");
419 printf ("const FcChar16 langBankNumbers[%d] = {\n",
422 for (i
= 0; sets
[i
]; i
++)
426 if (duplicate
[i
] >= 0)
428 for (n
= 0; n
< sets
[i
]->num
; n
++)
432 printf (" 0x%04x,", FcCharSetGetNumbers(sets
[i
])[n
]);
445 printf ("const FcLangCharSet fcLangCharSets[] = {\n");
446 for (i
= 0; sets
[i
]; i
++)
448 int j
= duplicate
[i
];
453 printf (" { (FcChar8 *) \"%s\",\n"
454 " { FC_REF_CONSTANT, %d, FC_BANK_LANGS, "
455 "{ { %d, %d } } } }, /* %d */\n",
457 sets
[j
]->num
, offsets
[j
], offsets
[j
], j
);
460 printf ("#define NUM_LANG_CHAR_SET %d\n", i
);
461 num_lang_set_map
= (i
+ 31) / 32;
462 printf ("#define NUM_LANG_SET_MAP %d\n", num_lang_set_map
);
464 * Dump indices with country codes
469 int ncountry_ent
= 0;
471 printf ("static const FcChar32 fcLangCountrySets[][NUM_LANG_SET_MAP] = {\n");
472 for (c
= 0; c
< ncountry
; c
++)
477 int lang
= strchr (langs
[i
], '-') - langs
[i
];
480 for (k
= 0; k
< num_lang_set_map
; k
++)
484 for (d
= c
+ 1; d
< ncountry
; d
++)
487 if (j
>= 0 && !strncmp (langs
[j
], langs
[i
], l
))
494 for (k
= 0; k
< num_lang_set_map
; k
++)
495 printf (" 0x%08x,", map
[k
]);
496 printf (" }, /* %*.*s */\n",
497 lang
, lang
, langs
[i
]);
502 printf ("#define NUM_COUNTRY_SET %d\n", ncountry_ent
);
507 * Dump sets start/finish for the fastpath
509 printf ("static const FcLangCharSetRange fcLangCharSetRanges[] = {\n");
510 for (setRangeChar
= 'a'; setRangeChar
<= 'z' ; setRangeChar
++)
512 printf (" { %d, %d }, /* %c */\n",
513 setRangeStart
[setRangeChar
- 'a'],
514 setRangeEnd
[setRangeChar
- 'a'], setRangeChar
);
518 while (fgets (line
, sizeof (line
), stdin
))
519 fputs (line
, stdout
);
522 exit (ferror (stdout
));