Italian translation update.
[rox-filer.git] / ROX-Filer / src / xdgmimecache.c
blobd5eae935f64d1d7d31d30e544311ae518d6ae1e0
1 /* -*- mode: C; c-file-style: "gnu" -*- */
2 /* xdgmimealias.c: Private file. mmappable caches for mime data
4 * More info can be found at http://www.freedesktop.org/standards/
6 * Copyright (C) 2005 Matthias Clasen <mclasen@redhat.com>
8 * Licensed under the Academic Free License version 2.0
9 * Or under the following terms:
11 * This library is free software; you can redistribute it and/or
12 * modify it under the terms of the GNU Lesser General Public
13 * License as published by the Free Software Foundation; either
14 * version 2 of the License, or (at your option) any later version.
16 * This library is distributed in the hope that it will be useful,
17 * but WITHOUT ANY WARRANTY; without even the implied warranty of
18 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
19 * Lesser General Public License for more details.
21 * You should have received a copy of the GNU Lesser General Public
22 * License along with this library; if not, write to the
23 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
24 * Boston, MA 02111-1307, USA.
27 #ifdef HAVE_CONFIG_H
28 #include <config.h>
29 #endif
31 #include <stdio.h>
32 #include <stdlib.h>
33 #include <string.h>
35 #include <fcntl.h>
36 #include <unistd.h>
37 #include <fnmatch.h>
38 #include <assert.h>
40 #include <netinet/in.h> /* for ntohl/ntohs */
42 #ifdef HAVE_MMAP
43 #include <sys/mman.h>
44 #else
45 #warning Building xdgmime without MMAP support. Binary "mime.info" cache files will not be used.
46 #endif
48 #include <sys/stat.h>
49 #include <sys/types.h>
51 #include "xdgmimecache.h"
52 #include "xdgmimeint.h"
54 #ifndef MAX
55 #define MAX(a,b) ((a) > (b) ? (a) : (b))
56 #endif
58 #ifndef FALSE
59 #define FALSE (0)
60 #endif
62 #ifndef TRUE
63 #define TRUE (!FALSE)
64 #endif
66 #ifndef _O_BINARY
67 #define _O_BINARY 0
68 #endif
70 #ifndef MAP_FAILED
71 #define MAP_FAILED ((void *) -1)
72 #endif
74 #define MAJOR_VERSION 1
75 #define MINOR_VERSION 0
77 struct _XdgMimeCache
79 int ref_count;
81 size_t size;
82 char *buffer;
85 #define GET_UINT16(cache,offset) (ntohs(*(xdg_uint16_t*)((cache) + (offset))))
86 #define GET_UINT32(cache,offset) (ntohl(*(xdg_uint32_t*)((cache) + (offset))))
88 XdgMimeCache *
89 _xdg_mime_cache_ref (XdgMimeCache *cache)
91 cache->ref_count++;
92 return cache;
95 void
96 _xdg_mime_cache_unref (XdgMimeCache *cache)
98 cache->ref_count--;
100 if (cache->ref_count == 0)
102 #ifdef HAVE_MMAP
103 munmap (cache->buffer, cache->size);
104 #endif
105 free (cache);
109 XdgMimeCache *
110 _xdg_mime_cache_new_from_file (const char *file_name)
112 XdgMimeCache *cache = NULL;
114 #ifdef HAVE_MMAP
115 int fd = -1;
116 struct stat st;
117 char *buffer = NULL;
119 /* Open the file and map it into memory */
120 fd = open (file_name, O_RDONLY|_O_BINARY, 0);
122 if (fd < 0)
123 return NULL;
125 if (fstat (fd, &st) < 0 || st.st_size < 4)
126 goto done;
128 buffer = (char *) mmap (NULL, st.st_size, PROT_READ, MAP_SHARED, fd, 0);
130 if (buffer == MAP_FAILED)
131 goto done;
133 /* Verify version */
134 if (GET_UINT16 (buffer, 0) != MAJOR_VERSION ||
135 GET_UINT16 (buffer, 2) != MINOR_VERSION)
137 munmap (buffer, st.st_size);
139 goto done;
142 cache = (XdgMimeCache *) malloc (sizeof (XdgMimeCache));
143 cache->ref_count = 1;
144 cache->buffer = buffer;
145 cache->size = st.st_size;
147 done:
148 if (fd != -1)
149 close (fd);
151 #endif /* HAVE_MMAP */
153 return cache;
156 static int
157 cache_magic_matchlet_compare_to_data (XdgMimeCache *cache,
158 xdg_uint32_t offset,
159 const void *data,
160 size_t len)
162 xdg_uint32_t range_start = GET_UINT32 (cache->buffer, offset);
163 xdg_uint32_t range_length = GET_UINT32 (cache->buffer, offset + 4);
164 xdg_uint32_t data_length = GET_UINT32 (cache->buffer, offset + 12);
165 xdg_uint32_t data_offset = GET_UINT32 (cache->buffer, offset + 16);
166 xdg_uint32_t mask_offset = GET_UINT32 (cache->buffer, offset + 20);
168 int i, j;
170 for (i = range_start; i <= range_start + range_length; i++)
172 int valid_matchlet = TRUE;
174 if (i + data_length > len)
175 return FALSE;
177 if (mask_offset)
179 for (j = 0; j < data_length; j++)
181 if ((((unsigned char *)cache->buffer)[data_offset + j] & ((unsigned char *)cache->buffer)[mask_offset + j]) !=
182 ((((unsigned char *) data)[j + i]) & ((unsigned char *)cache->buffer)[mask_offset + j]))
184 valid_matchlet = FALSE;
185 break;
189 else
191 for (j = 0; j < data_length; j++)
193 if (((unsigned char *)cache->buffer)[data_offset + j] != ((unsigned char *) data)[j + i])
195 valid_matchlet = FALSE;
196 break;
201 if (valid_matchlet)
202 return TRUE;
205 return FALSE;
208 static int
209 cache_magic_matchlet_compare (XdgMimeCache *cache,
210 xdg_uint32_t offset,
211 const void *data,
212 size_t len)
214 xdg_uint32_t n_children = GET_UINT32 (cache->buffer, offset + 24);
215 xdg_uint32_t child_offset = GET_UINT32 (cache->buffer, offset + 28);
217 int i;
219 if (cache_magic_matchlet_compare_to_data (cache, offset, data, len))
221 if (n_children == 0)
222 return TRUE;
224 for (i = 0; i < n_children; i++)
226 if (cache_magic_matchlet_compare (cache, child_offset + 32 * i,
227 data, len))
228 return TRUE;
232 return FALSE;
235 static const char *
236 cache_magic_compare_to_data (XdgMimeCache *cache,
237 xdg_uint32_t offset,
238 const void *data,
239 size_t len,
240 int *prio)
242 xdg_uint32_t priority = GET_UINT32 (cache->buffer, offset);
243 xdg_uint32_t mimetype_offset = GET_UINT32 (cache->buffer, offset + 4);
244 xdg_uint32_t n_matchlets = GET_UINT32 (cache->buffer, offset + 8);
245 xdg_uint32_t matchlet_offset = GET_UINT32 (cache->buffer, offset + 12);
247 int i;
249 for (i = 0; i < n_matchlets; i++)
251 if (cache_magic_matchlet_compare (cache, matchlet_offset + i * 32,
252 data, len))
254 *prio = priority;
256 return cache->buffer + mimetype_offset;
260 return NULL;
263 static const char *
264 cache_magic_lookup_data (XdgMimeCache *cache,
265 const void *data,
266 size_t len,
267 int *prio,
268 const char *mime_types[],
269 int n_mime_types)
271 xdg_uint32_t list_offset;
272 xdg_uint32_t n_entries;
273 xdg_uint32_t offset;
275 int j, n;
277 *prio = 0;
279 list_offset = GET_UINT32 (cache->buffer, 24);
280 n_entries = GET_UINT32 (cache->buffer, list_offset);
281 offset = GET_UINT32 (cache->buffer, list_offset + 8);
283 for (j = 0; j < n_entries; j++)
285 const char *match;
287 match = cache_magic_compare_to_data (cache, offset + 16 * j,
288 data, len, prio);
289 if (match)
290 return match;
291 else
293 xdg_uint32_t mimetype_offset;
294 const char *non_match;
296 mimetype_offset = GET_UINT32 (cache->buffer, offset + 16 * j + 4);
297 non_match = cache->buffer + mimetype_offset;
299 for (n = 0; n < n_mime_types; n++)
301 if (mime_types[n] &&
302 xdg_mime_mime_type_equal (mime_types[n], non_match))
303 mime_types[n] = NULL;
308 return NULL;
311 static const char *
312 cache_alias_lookup (const char *alias)
314 const char *ptr;
315 int i, min, max, mid, cmp;
317 for (i = 0; _xdg_mime_caches[i]; i++)
319 XdgMimeCache *cache = _xdg_mime_caches[i];
320 xdg_uint32_t list_offset = GET_UINT32 (cache->buffer, 4);
321 xdg_uint32_t n_entries = GET_UINT32 (cache->buffer, list_offset);
322 xdg_uint32_t offset;
324 min = 0;
325 max = n_entries - 1;
326 while (max >= min)
328 mid = (min + max) / 2;
330 offset = GET_UINT32 (cache->buffer, list_offset + 4 + 8 * mid);
331 ptr = cache->buffer + offset;
332 cmp = strcmp (ptr, alias);
334 if (cmp < 0)
335 min = mid + 1;
336 else if (cmp > 0)
337 max = mid - 1;
338 else
340 offset = GET_UINT32 (cache->buffer, list_offset + 4 + 8 * mid + 4);
341 return cache->buffer + offset;
346 return NULL;
349 static int
350 cache_glob_lookup_literal (const char *file_name,
351 const char *mime_types[],
352 int n_mime_types)
354 const char *ptr;
355 int i, min, max, mid, cmp;
357 for (i = 0; _xdg_mime_caches[i]; i++)
359 XdgMimeCache *cache = _xdg_mime_caches[i];
360 xdg_uint32_t list_offset = GET_UINT32 (cache->buffer, 12);
361 xdg_uint32_t n_entries = GET_UINT32 (cache->buffer, list_offset);
362 xdg_uint32_t offset;
364 min = 0;
365 max = n_entries - 1;
366 while (max >= min)
368 mid = (min + max) / 2;
370 offset = GET_UINT32 (cache->buffer, list_offset + 4 + 8 * mid);
371 ptr = cache->buffer + offset;
372 cmp = strcmp (ptr, file_name);
374 if (cmp < 0)
375 min = mid + 1;
376 else if (cmp > 0)
377 max = mid - 1;
378 else
380 offset = GET_UINT32 (cache->buffer, list_offset + 4 + 8 * mid + 4);
381 mime_types[0] = (const char *)(cache->buffer + offset);
383 return 1;
388 return 0;
391 static int
392 cache_glob_lookup_fnmatch (const char *file_name,
393 const char *mime_types[],
394 int n_mime_types)
396 const char *mime_type;
397 const char *ptr;
399 int i, j, n;
401 n = 0;
402 for (i = 0; _xdg_mime_caches[i]; i++)
404 XdgMimeCache *cache = _xdg_mime_caches[i];
406 xdg_uint32_t list_offset = GET_UINT32 (cache->buffer, 20);
407 xdg_uint32_t n_entries = GET_UINT32 (cache->buffer, list_offset);
409 for (j = 0; j < n_entries && n < n_mime_types; j++)
411 xdg_uint32_t offset = GET_UINT32 (cache->buffer, list_offset + 4 + 8 * j);
412 xdg_uint32_t mimetype_offset = GET_UINT32 (cache->buffer, list_offset + 4 + 8 * j + 4);
413 ptr = cache->buffer + offset;
414 mime_type = cache->buffer + mimetype_offset;
416 /* FIXME: Not UTF-8 safe */
417 if (fnmatch (ptr, file_name, 0) == 0)
418 mime_types[n++] = mime_type;
421 if (n > 0)
422 return n;
425 return 0;
428 static int
429 cache_glob_node_lookup_suffix (XdgMimeCache *cache,
430 xdg_uint32_t n_entries,
431 xdg_uint32_t offset,
432 const char *suffix,
433 int ignore_case,
434 const char *mime_types[],
435 int n_mime_types)
437 xdg_unichar_t character;
438 xdg_unichar_t match_char;
439 xdg_uint32_t mimetype_offset;
440 xdg_uint32_t n_children;
441 xdg_uint32_t child_offset;
443 int min, max, mid, n, i;
445 character = _xdg_utf8_to_ucs4 (suffix);
446 if (ignore_case)
447 character = _xdg_ucs4_to_lower (character);
449 min = 0;
450 max = n_entries - 1;
451 while (max >= min)
453 mid = (min + max) / 2;
455 match_char = GET_UINT32 (cache->buffer, offset + 16 * mid);
457 if (match_char < character)
458 min = mid + 1;
459 else if (match_char > character)
460 max = mid - 1;
461 else
463 suffix = _xdg_utf8_next_char (suffix);
464 if (*suffix == '\0')
466 mimetype_offset = GET_UINT32 (cache->buffer, offset + 16 * mid + 4);
467 n = 0;
468 if (cache->buffer[mimetype_offset])
469 mime_types[n++] = cache->buffer + mimetype_offset;
471 n_children = GET_UINT32 (cache->buffer, offset + 16 * mid + 8);
472 child_offset = GET_UINT32 (cache->buffer, offset + 16 * mid + 12);
473 i = 0;
474 while (n < n_mime_types && i < n_children)
476 match_char = GET_UINT32 (cache->buffer, child_offset + 16 * i);
477 mimetype_offset = GET_UINT32 (cache->buffer, offset + 16 * i + 4);
478 if (match_char != 0)
479 break;
481 mime_types[n++] = cache->buffer + mimetype_offset;
482 i++;
485 return n;
487 else
489 n_children = GET_UINT32 (cache->buffer, offset + 16 * mid + 8);
490 child_offset = GET_UINT32 (cache->buffer, offset + 16 * mid + 12);
492 return cache_glob_node_lookup_suffix (cache,
493 n_children, child_offset,
494 suffix, ignore_case,
495 mime_types,
496 n_mime_types);
501 return 0;
504 static int
505 cache_glob_lookup_suffix (const char *suffix,
506 int ignore_case,
507 const char *mime_types[],
508 int n_mime_types)
510 int i, n;
512 for (i = 0; _xdg_mime_caches[i]; i++)
514 XdgMimeCache *cache = _xdg_mime_caches[i];
516 xdg_uint32_t list_offset = GET_UINT32 (cache->buffer, 16);
517 xdg_uint32_t n_entries = GET_UINT32 (cache->buffer, list_offset);
518 xdg_uint32_t offset = GET_UINT32 (cache->buffer, list_offset + 4);
520 n = cache_glob_node_lookup_suffix (cache,
521 n_entries, offset,
522 suffix, ignore_case,
523 mime_types,
524 n_mime_types);
525 if (n > 0)
526 return n;
529 return 0;
532 static void
533 find_stopchars (char *stopchars)
535 int i, j, k, l;
537 k = 0;
538 for (i = 0; _xdg_mime_caches[i]; i++)
540 XdgMimeCache *cache = _xdg_mime_caches[i];
542 xdg_uint32_t list_offset = GET_UINT32 (cache->buffer, 16);
543 xdg_uint32_t n_entries = GET_UINT32 (cache->buffer, list_offset);
544 xdg_uint32_t offset = GET_UINT32 (cache->buffer, list_offset + 4);
546 for (j = 0; j < n_entries; j++)
548 xdg_uint32_t match_char = GET_UINT32 (cache->buffer, offset);
550 if (match_char < 128)
552 for (l = 0; l < k; l++)
553 if (stopchars[l] == match_char)
554 break;
555 if (l == k)
557 stopchars[k] = (char) match_char;
558 k++;
562 offset += 16;
566 stopchars[k] = '\0';
569 static int
570 cache_glob_lookup_file_name (const char *file_name,
571 const char *mime_types[],
572 int n_mime_types)
574 const char *ptr;
575 char stopchars[128];
576 int n;
578 assert (file_name != NULL);
580 /* First, check the literals */
581 n = cache_glob_lookup_literal (file_name, mime_types, n_mime_types);
582 if (n > 0)
583 return n;
585 find_stopchars (stopchars);
587 /* Next, check suffixes */
588 ptr = strpbrk (file_name, stopchars);
589 while (ptr)
591 n = cache_glob_lookup_suffix (ptr, FALSE, mime_types, n_mime_types);
592 if (n > 0)
593 return n;
595 n = cache_glob_lookup_suffix (ptr, TRUE, mime_types, n_mime_types);
596 if (n > 0)
597 return n;
599 ptr = strpbrk (ptr + 1, stopchars);
602 /* Last, try fnmatch */
603 return cache_glob_lookup_fnmatch (file_name, mime_types, n_mime_types);
607 _xdg_mime_cache_get_max_buffer_extents (void)
609 xdg_uint32_t offset;
610 xdg_uint32_t max_extent;
611 int i;
613 max_extent = 0;
614 for (i = 0; _xdg_mime_caches[i]; i++)
616 XdgMimeCache *cache = _xdg_mime_caches[i];
618 offset = GET_UINT32 (cache->buffer, 24);
619 max_extent = MAX (max_extent, GET_UINT32 (cache->buffer, offset + 4));
622 return max_extent;
625 static const char *
626 cache_get_mime_type_for_data (const void *data,
627 size_t len,
628 const char *mime_types[],
629 int n_mime_types)
631 const char *mime_type;
632 int i, n, priority;
634 priority = 0;
635 mime_type = NULL;
636 for (i = 0; _xdg_mime_caches[i]; i++)
638 XdgMimeCache *cache = _xdg_mime_caches[i];
640 int prio;
641 const char *match;
643 match = cache_magic_lookup_data (cache, data, len, &prio,
644 mime_types, n_mime_types);
645 if (prio > priority)
647 priority = prio;
648 mime_type = match;
652 if (priority > 0)
653 return mime_type;
655 for (n = 0; n < n_mime_types; n++)
657 if (mime_types[n])
658 return mime_types[n];
661 return XDG_MIME_TYPE_UNKNOWN;
664 const char *
665 _xdg_mime_cache_get_mime_type_for_data (const void *data,
666 size_t len)
668 return cache_get_mime_type_for_data (data, len, NULL, 0);
671 const char *
672 _xdg_mime_cache_get_mime_type_for_file (const char *file_name,
673 struct stat *statbuf)
675 const char *mime_type;
676 const char *mime_types[2];
677 FILE *file;
678 unsigned char *data;
679 int max_extent;
680 int bytes_read;
681 struct stat buf;
682 const char *base_name;
683 int n;
685 if (file_name == NULL)
686 return NULL;
688 if (! _xdg_utf8_validate (file_name))
689 return NULL;
691 base_name = _xdg_get_base_name (file_name);
692 n = cache_glob_lookup_file_name (base_name, mime_types, 2);
694 if (n == 1)
695 return mime_types[0];
697 if (!statbuf)
699 if (stat (file_name, &buf) != 0)
700 return XDG_MIME_TYPE_UNKNOWN;
702 statbuf = &buf;
705 if (!S_ISREG (statbuf->st_mode))
706 return XDG_MIME_TYPE_UNKNOWN;
708 /* FIXME: Need to make sure that max_extent isn't totally broken. This could
709 * be large and need getting from a stream instead of just reading it all
710 * in. */
711 max_extent = _xdg_mime_cache_get_max_buffer_extents ();
712 data = malloc (max_extent);
713 if (data == NULL)
714 return XDG_MIME_TYPE_UNKNOWN;
716 file = fopen (file_name, "r");
717 if (file == NULL)
719 free (data);
720 return XDG_MIME_TYPE_UNKNOWN;
723 bytes_read = fread (data, 1, max_extent, file);
724 if (ferror (file))
726 free (data);
727 fclose (file);
728 return XDG_MIME_TYPE_UNKNOWN;
731 mime_type = cache_get_mime_type_for_data (data, bytes_read,
732 mime_types, n);
734 free (data);
735 fclose (file);
737 return mime_type;
740 const char *
741 _xdg_mime_cache_get_mime_type_from_file_name (const char *file_name)
743 const char *mime_types[2];
745 if (cache_glob_lookup_file_name (file_name, mime_types, 2) == 1)
746 return mime_types[0];
747 else
748 return XDG_MIME_TYPE_UNKNOWN;
751 #if 1
752 static int
753 is_super_type (const char *mime)
755 int length;
756 const char *type;
758 length = strlen (mime);
759 type = &(mime[length - 2]);
761 if (strcmp (type, "/*") == 0)
762 return 1;
764 return 0;
766 #endif
769 _xdg_mime_cache_mime_type_subclass (const char *mime,
770 const char *base)
772 const char *umime, *ubase;
774 int i, j, min, max, med, cmp;
776 umime = _xdg_mime_cache_unalias_mime_type (mime);
777 ubase = _xdg_mime_cache_unalias_mime_type (base);
779 if (strcmp (umime, ubase) == 0)
780 return 1;
782 /* We really want to handle text/ * in GtkFileFilter, so we just
783 * turn on the supertype matching
785 #if 1
786 /* Handle supertypes */
787 if (is_super_type (ubase) &&
788 xdg_mime_media_type_equal (umime, ubase))
789 return 1;
790 #endif
792 /* Handle special cases text/plain and application/octet-stream */
793 if (strcmp (ubase, "text/plain") == 0 &&
794 strncmp (umime, "text/", 5) == 0)
795 return 1;
797 if (strcmp (ubase, "application/octet-stream") == 0)
798 return 1;
800 for (i = 0; _xdg_mime_caches[i]; i++)
802 XdgMimeCache *cache = _xdg_mime_caches[i];
804 xdg_uint32_t list_offset = GET_UINT32 (cache->buffer, 8);
805 xdg_uint32_t n_entries = GET_UINT32 (cache->buffer, list_offset);
806 xdg_uint32_t offset, n_parents, parent_offset;
808 min = 0;
809 max = n_entries - 1;
810 while (max >= min)
812 med = (min + max)/2;
814 offset = GET_UINT32 (cache->buffer, list_offset + 4 + 8 * med);
815 cmp = strcmp (cache->buffer + offset, umime);
816 if (cmp < 0)
817 min = med + 1;
818 else if (cmp > 0)
819 max = med - 1;
820 else
822 offset = GET_UINT32 (cache->buffer, list_offset + 4 + 8 * med + 4);
823 n_parents = GET_UINT32 (cache->buffer, offset);
825 for (j = 0; j < n_parents; j++)
827 parent_offset = GET_UINT32 (cache->buffer, offset + 4 + 4 * j);
828 if (_xdg_mime_cache_mime_type_subclass (cache->buffer + parent_offset, ubase))
829 return 1;
832 break;
837 return 0;
840 const char *
841 _xdg_mime_cache_unalias_mime_type (const char *mime)
843 const char *lookup;
845 lookup = cache_alias_lookup (mime);
847 if (lookup)
848 return lookup;
850 return mime;
853 char **
854 _xdg_mime_cache_list_mime_parents (const char *mime)
856 int i, j, p;
857 char *all_parents[128]; /* we'll stop at 128 */
858 char **result;
860 mime = xdg_mime_unalias_mime_type (mime);
862 p = 0;
863 for (i = 0; _xdg_mime_caches[i]; i++)
865 XdgMimeCache *cache = _xdg_mime_caches[i];
867 xdg_uint32_t list_offset = GET_UINT32 (cache->buffer, 8);
868 xdg_uint32_t n_entries = GET_UINT32 (cache->buffer, list_offset);
870 for (j = 0; j < n_entries; j++)
872 xdg_uint32_t mimetype_offset = GET_UINT32 (cache->buffer, list_offset + 4 + 8 * j);
873 xdg_uint32_t parents_offset = GET_UINT32 (cache->buffer, list_offset + 4 + 8 * j + 4);
875 if (strcmp (cache->buffer + mimetype_offset, mime) == 0)
877 int k;
878 xdg_uint32_t parent_mime_offset;
879 xdg_uint32_t n_parents = GET_UINT32 (cache->buffer, parents_offset);
881 for (k = 0; k < n_parents && p < 127; k++)
883 parent_mime_offset = GET_UINT32 (cache->buffer, parents_offset + 4 + 4 * k);
884 all_parents[p++] = cache->buffer + parent_mime_offset;
887 break;
891 all_parents[p++] = 0;
893 result = (char **) malloc (p * sizeof (char *));
894 memcpy (result, all_parents, p * sizeof (char *));
896 return result;