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.
40 #include <netinet/in.h> /* for ntohl/ntohs */
45 #warning Building xdgmime without MMAP support. Binary "mime.cache" files will not be used.
49 #include <sys/types.h>
51 #include "xdgmimecache.h"
52 #include "xdgmimeint.h"
55 #define MAX(a,b) ((a) > (b) ? (a) : (b))
71 #define MAP_FAILED ((void *) -1)
74 #define MAJOR_VERSION 1
75 #define MINOR_VERSION_MIN 1
76 #define MINOR_VERSION_MAX 2
87 #define GET_UINT16(cache,offset) (ntohs(*(xdg_uint16_t*)((cache) + (offset))))
88 #define GET_UINT32(cache,offset) (ntohl(*(xdg_uint32_t*)((cache) + (offset))))
91 _xdg_mime_cache_ref (XdgMimeCache
*cache
)
98 _xdg_mime_cache_unref (XdgMimeCache
*cache
)
102 if (cache
->ref_count
== 0)
105 munmap (cache
->buffer
, cache
->size
);
112 _xdg_mime_cache_new_from_file (const char *file_name
)
114 XdgMimeCache
*cache
= NULL
;
122 /* Open the file and map it into memory */
123 fd
= open (file_name
, O_RDONLY
|_O_BINARY
, 0);
128 if (fstat (fd
, &st
) < 0 || st
.st_size
< 4)
131 buffer
= (char *) mmap (NULL
, st
.st_size
, PROT_READ
, MAP_SHARED
, fd
, 0);
133 if (buffer
== MAP_FAILED
)
136 minor
= GET_UINT16 (buffer
, 2);
138 if (GET_UINT16 (buffer
, 0) != MAJOR_VERSION
||
139 (minor
< MINOR_VERSION_MIN
||
140 minor
> MINOR_VERSION_MAX
))
142 munmap (buffer
, st
.st_size
);
147 cache
= (XdgMimeCache
*) malloc (sizeof (XdgMimeCache
));
148 cache
->minor
= minor
;
149 cache
->ref_count
= 1;
150 cache
->buffer
= buffer
;
151 cache
->size
= st
.st_size
;
157 #endif /* HAVE_MMAP */
163 cache_magic_matchlet_compare_to_data (XdgMimeCache
*cache
,
168 xdg_uint32_t range_start
= GET_UINT32 (cache
->buffer
, offset
);
169 xdg_uint32_t range_length
= GET_UINT32 (cache
->buffer
, offset
+ 4);
170 xdg_uint32_t data_length
= GET_UINT32 (cache
->buffer
, offset
+ 12);
171 xdg_uint32_t data_offset
= GET_UINT32 (cache
->buffer
, offset
+ 16);
172 xdg_uint32_t mask_offset
= GET_UINT32 (cache
->buffer
, offset
+ 20);
176 for (i
= range_start
; i
< range_start
+ range_length
; i
++)
178 int valid_matchlet
= TRUE
;
180 if (i
+ data_length
> len
)
185 for (j
= 0; j
< data_length
; j
++)
187 if ((((unsigned char *)cache
->buffer
)[data_offset
+ j
] & ((unsigned char *)cache
->buffer
)[mask_offset
+ j
]) !=
188 ((((unsigned char *) data
)[j
+ i
]) & ((unsigned char *)cache
->buffer
)[mask_offset
+ j
]))
190 valid_matchlet
= FALSE
;
197 valid_matchlet
= memcmp(cache
->buffer
+ data_offset
, data
+ i
, data_length
) == 0;
208 cache_magic_matchlet_compare (XdgMimeCache
*cache
,
213 xdg_uint32_t n_children
= GET_UINT32 (cache
->buffer
, offset
+ 24);
214 xdg_uint32_t child_offset
= GET_UINT32 (cache
->buffer
, offset
+ 28);
218 if (cache_magic_matchlet_compare_to_data (cache
, offset
, data
, len
))
223 for (i
= 0; i
< n_children
; i
++)
225 if (cache_magic_matchlet_compare (cache
, child_offset
+ 32 * i
,
235 cache_magic_compare_to_data (XdgMimeCache
*cache
,
241 xdg_uint32_t priority
= GET_UINT32 (cache
->buffer
, offset
);
242 xdg_uint32_t mimetype_offset
= GET_UINT32 (cache
->buffer
, offset
+ 4);
243 xdg_uint32_t n_matchlets
= GET_UINT32 (cache
->buffer
, offset
+ 8);
244 xdg_uint32_t matchlet_offset
= GET_UINT32 (cache
->buffer
, offset
+ 12);
248 for (i
= 0; i
< n_matchlets
; i
++)
250 if (cache_magic_matchlet_compare (cache
, matchlet_offset
+ i
* 32,
255 return cache
->buffer
+ mimetype_offset
;
263 cache_magic_lookup_data (XdgMimeCache
*cache
,
267 const char *mime_types
[],
270 xdg_uint32_t list_offset
;
271 xdg_uint32_t n_entries
;
278 list_offset
= GET_UINT32 (cache
->buffer
, 24);
279 n_entries
= GET_UINT32 (cache
->buffer
, list_offset
);
280 offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 8);
282 for (j
= 0; j
< n_entries
; j
++)
286 match
= cache_magic_compare_to_data (cache
, offset
+ 16 * j
,
292 xdg_uint32_t mimetype_offset
;
293 const char *non_match
;
295 mimetype_offset
= GET_UINT32 (cache
->buffer
, offset
+ 16 * j
+ 4);
296 non_match
= cache
->buffer
+ mimetype_offset
;
298 for (n
= 0; n
< n_mime_types
; n
++)
301 _xdg_mime_mime_type_equal (mime_types
[n
], non_match
))
302 mime_types
[n
] = NULL
;
311 cache_alias_lookup (const char *alias
)
314 int i
, min
, max
, mid
, cmp
;
316 for (i
= 0; _caches
[i
]; i
++)
318 XdgMimeCache
*cache
= _caches
[i
];
319 xdg_uint32_t list_offset
= GET_UINT32 (cache
->buffer
, 4);
320 xdg_uint32_t n_entries
= GET_UINT32 (cache
->buffer
, list_offset
);
327 mid
= (min
+ max
) / 2;
329 offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 8 * mid
);
330 ptr
= cache
->buffer
+ offset
;
331 cmp
= strcmp (ptr
, alias
);
339 offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 8 * mid
+ 4);
340 return cache
->buffer
+ offset
;
354 cache_glob_lookup_literal (const char *file_name
,
355 const char *mime_types
[],
357 int case_sensitive_check
)
360 int i
, min
, max
, mid
, cmp
;
362 for (i
= 0; _caches
[i
]; i
++)
364 XdgMimeCache
*cache
= _caches
[i
];
365 xdg_uint32_t list_offset
= GET_UINT32 (cache
->buffer
, 12);
366 xdg_uint32_t n_entries
= GET_UINT32 (cache
->buffer
, list_offset
);
373 mid
= (min
+ max
) / 2;
375 offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 12 * mid
);
376 ptr
= cache
->buffer
+ offset
;
377 cmp
= strcmp (ptr
, file_name
);
385 int weight
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 12 * mid
+ 8);
386 int case_sensitive
= weight
& 0x100;
387 weight
= weight
& 0xff;
389 if (case_sensitive_check
|| !case_sensitive
)
391 offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 12 * mid
+ 4);
392 mime_types
[0] = (const char *)(cache
->buffer
+ offset
);
405 cache_glob_lookup_fnmatch (const char *file_name
,
406 MimeWeight mime_types
[],
408 int case_sensitive_check
)
410 const char *mime_type
;
416 for (i
= 0; _caches
[i
]; i
++)
418 XdgMimeCache
*cache
= _caches
[i
];
420 xdg_uint32_t list_offset
= GET_UINT32 (cache
->buffer
, 20);
421 xdg_uint32_t n_entries
= GET_UINT32 (cache
->buffer
, list_offset
);
423 for (j
= 0; j
< n_entries
&& n
< n_mime_types
; j
++)
425 xdg_uint32_t offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 12 * j
);
426 xdg_uint32_t mimetype_offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 12 * j
+ 4);
427 int weight
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 12 * j
+ 8);
428 int case_sensitive
= weight
& 0x100;
429 weight
= weight
& 0xff;
430 ptr
= cache
->buffer
+ offset
;
431 mime_type
= cache
->buffer
+ mimetype_offset
;
432 if (case_sensitive_check
|| !case_sensitive
)
434 /* FIXME: Not UTF-8 safe */
435 if (fnmatch (ptr
, file_name
, 0) == 0)
437 mime_types
[n
].mime
= mime_type
;
438 mime_types
[n
].weight
= weight
;
452 cache_glob_node_lookup_suffix (XdgMimeCache
*cache
,
453 xdg_uint32_t n_entries
,
455 const char *file_name
,
457 int case_sensitive_check
,
458 MimeWeight mime_types
[],
461 xdg_unichar_t character
;
462 xdg_unichar_t match_char
;
463 xdg_uint32_t mimetype_offset
;
464 xdg_uint32_t n_children
;
465 xdg_uint32_t child_offset
;
469 int min
, max
, mid
, n
, i
;
471 character
= file_name
[len
- 1];
473 assert (character
!= 0);
479 mid
= (min
+ max
) / 2;
480 match_char
= GET_UINT32 (cache
->buffer
, offset
+ 12 * mid
);
481 if (match_char
< character
)
483 else if (match_char
> character
)
489 n_children
= GET_UINT32 (cache
->buffer
, offset
+ 12 * mid
+ 4);
490 child_offset
= GET_UINT32 (cache
->buffer
, offset
+ 12 * mid
+ 8);
494 n
= cache_glob_node_lookup_suffix (cache
,
495 n_children
, child_offset
,
497 case_sensitive_check
,
504 while (n
< n_mime_types
&& i
< n_children
)
506 match_char
= GET_UINT32 (cache
->buffer
, child_offset
+ 12 * i
);
510 mimetype_offset
= GET_UINT32 (cache
->buffer
, child_offset
+ 12 * i
+ 4);
511 weight
= GET_UINT32 (cache
->buffer
, child_offset
+ 12 * i
+ 8);
512 case_sensitive
= weight
& 0x100;
513 weight
= weight
& 0xff;
515 if (case_sensitive_check
|| !case_sensitive
)
517 mime_types
[n
].mime
= cache
->buffer
+ mimetype_offset
;
518 mime_types
[n
].weight
= weight
;
531 cache_glob_lookup_suffix (const char *file_name
,
534 MimeWeight mime_types
[],
539 for (i
= 0; _caches
[i
]; i
++)
541 XdgMimeCache
*cache
= _caches
[i
];
543 xdg_uint32_t list_offset
= GET_UINT32 (cache
->buffer
, 16);
544 xdg_uint32_t n_entries
= GET_UINT32 (cache
->buffer
, list_offset
);
545 xdg_uint32_t offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4);
547 n
= cache_glob_node_lookup_suffix (cache
,
560 static int compare_mime_weight (const void *a
, const void *b
)
562 const MimeWeight
*aa
= (const MimeWeight
*)a
;
563 const MimeWeight
*bb
= (const MimeWeight
*)b
;
565 return bb
->weight
- aa
->weight
;
568 #define ISUPPER(c) ((c) >= 'A' && (c) <= 'Z')
570 ascii_tolower (const char *str
)
574 lower
= strdup (str
);
579 *p
++ = ISUPPER (c
) ? c
- 'A' + 'a' : c
;
585 cache_glob_lookup_file_name (const char *file_name
,
586 const char *mime_types
[],
590 MimeWeight mimes
[10];
596 assert (file_name
!= NULL
&& n_mime_types
> 0);
598 /* First, check the literals */
600 lower_case
= ascii_tolower (file_name
);
602 n
= cache_glob_lookup_literal (lower_case
, mime_types
, n_mime_types
, FALSE
);
609 n
= cache_glob_lookup_literal (file_name
, mime_types
, n_mime_types
, TRUE
);
616 len
= strlen (file_name
);
617 n
= cache_glob_lookup_suffix (lower_case
, len
, FALSE
, mimes
, n_mimes
);
619 n
= cache_glob_lookup_suffix (file_name
, len
, TRUE
, mimes
, n_mimes
);
621 /* Last, try fnmatch */
623 n
= cache_glob_lookup_fnmatch (lower_case
, mimes
, n_mimes
, FALSE
);
625 n
= cache_glob_lookup_fnmatch (file_name
, mimes
, n_mimes
, TRUE
);
629 qsort (mimes
, n
, sizeof (MimeWeight
), compare_mime_weight
);
631 if (n_mime_types
< n
)
634 for (i
= 0; i
< n
; i
++)
635 mime_types
[i
] = mimes
[i
].mime
;
641 _xdg_mime_cache_get_max_buffer_extents (void)
644 xdg_uint32_t max_extent
;
648 for (i
= 0; _caches
[i
]; i
++)
650 XdgMimeCache
*cache
= _caches
[i
];
652 offset
= GET_UINT32 (cache
->buffer
, 24);
653 max_extent
= MAX (max_extent
, GET_UINT32 (cache
->buffer
, offset
+ 4));
660 cache_get_mime_type_for_data (const void *data
,
663 const char *mime_types
[],
666 const char *mime_type
;
671 for (i
= 0; _caches
[i
]; i
++)
673 XdgMimeCache
*cache
= _caches
[i
];
678 match
= cache_magic_lookup_data (cache
, data
, len
, &prio
,
679 mime_types
, n_mime_types
);
688 *result_prio
= priority
;
692 /* Pick glob-result R where mime_type inherits from R */
693 for (n
= 0; n
< n_mime_types
; n
++)
695 if (mime_types
[n
] && _xdg_mime_cache_mime_type_subclass(mime_types
[n
], mime_type
))
696 return mime_types
[n
];
699 /* Return magic match */
703 /* Pick first glob result, as fallback */
704 for (n
= 0; n
< n_mime_types
; n
++)
707 return mime_types
[n
];
714 _xdg_mime_cache_get_mime_type_for_data (const void *data
,
718 return cache_get_mime_type_for_data (data
, len
, result_prio
, NULL
, 0);
722 _xdg_mime_cache_get_mime_type_for_file (const char *file_name
,
723 struct stat
*statbuf
)
725 const char *mime_type
;
726 const char *mime_types
[10];
732 const char *base_name
;
735 if (file_name
== NULL
)
738 if (! _xdg_utf8_validate (file_name
))
741 base_name
= _xdg_get_base_name (file_name
);
742 n
= cache_glob_lookup_file_name (base_name
, mime_types
, 10);
745 return mime_types
[0];
749 if (stat (file_name
, &buf
) != 0)
750 return XDG_MIME_TYPE_UNKNOWN
;
755 if (statbuf
->st_size
== 0)
756 return XDG_MIME_TYPE_EMPTY
;
758 if (!S_ISREG (statbuf
->st_mode
))
759 return XDG_MIME_TYPE_UNKNOWN
;
761 /* FIXME: Need to make sure that max_extent isn't totally broken. This could
762 * be large and need getting from a stream instead of just reading it all
764 max_extent
= _xdg_mime_cache_get_max_buffer_extents ();
765 data
= malloc (max_extent
);
767 return XDG_MIME_TYPE_UNKNOWN
;
769 file
= fopen (file_name
, "r");
773 return XDG_MIME_TYPE_UNKNOWN
;
776 bytes_read
= fread (data
, 1, max_extent
, file
);
781 return XDG_MIME_TYPE_UNKNOWN
;
784 mime_type
= cache_get_mime_type_for_data (data
, bytes_read
, NULL
,
788 mime_type
= _xdg_binary_or_text_fallback(data
, bytes_read
);
797 _xdg_mime_cache_get_mime_type_from_file_name (const char *file_name
)
799 const char *mime_type
;
801 if (cache_glob_lookup_file_name (file_name
, &mime_type
, 1))
804 return XDG_MIME_TYPE_UNKNOWN
;
808 _xdg_mime_cache_get_mime_types_from_file_name (const char *file_name
,
809 const char *mime_types
[],
812 return cache_glob_lookup_file_name (file_name
, mime_types
, n_mime_types
);
817 is_super_type (const char *mime
)
822 length
= strlen (mime
);
823 type
= &(mime
[length
- 2]);
825 if (strcmp (type
, "/*") == 0)
833 _xdg_mime_cache_mime_type_subclass (const char *mime
,
836 const char *umime
, *ubase
;
838 int i
, j
, min
, max
, med
, cmp
;
840 umime
= _xdg_mime_cache_unalias_mime_type (mime
);
841 ubase
= _xdg_mime_cache_unalias_mime_type (base
);
843 if (strcmp (umime
, ubase
) == 0)
846 /* We really want to handle text/ * in GtkFileFilter, so we just
847 * turn on the supertype matching
850 /* Handle supertypes */
851 if (is_super_type (ubase
) &&
852 xdg_mime_media_type_equal (umime
, ubase
))
856 /* Handle special cases text/plain and application/octet-stream */
857 if (strcmp (ubase
, "text/plain") == 0 &&
858 strncmp (umime
, "text/", 5) == 0)
861 if (strcmp (ubase
, "application/octet-stream") == 0)
864 for (i
= 0; _caches
[i
]; i
++)
866 XdgMimeCache
*cache
= _caches
[i
];
868 xdg_uint32_t list_offset
= GET_UINT32 (cache
->buffer
, 8);
869 xdg_uint32_t n_entries
= GET_UINT32 (cache
->buffer
, list_offset
);
870 xdg_uint32_t offset
, n_parents
, parent_offset
;
878 offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 8 * med
);
879 cmp
= strcmp (cache
->buffer
+ offset
, umime
);
886 offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 8 * med
+ 4);
887 n_parents
= GET_UINT32 (cache
->buffer
, offset
);
889 for (j
= 0; j
< n_parents
; j
++)
891 parent_offset
= GET_UINT32 (cache
->buffer
, offset
+ 4 + 4 * j
);
892 if (_xdg_mime_cache_mime_type_subclass (cache
->buffer
+ parent_offset
, ubase
))
905 _xdg_mime_cache_unalias_mime_type (const char *mime
)
909 lookup
= cache_alias_lookup (mime
);
918 _xdg_mime_cache_list_mime_parents (const char *mime
)
921 char *all_parents
[128]; /* we'll stop at 128 */
924 mime
= xdg_mime_unalias_mime_type (mime
);
927 for (i
= 0; _caches
[i
]; i
++)
929 XdgMimeCache
*cache
= _caches
[i
];
931 xdg_uint32_t list_offset
= GET_UINT32 (cache
->buffer
, 8);
932 xdg_uint32_t n_entries
= GET_UINT32 (cache
->buffer
, list_offset
);
934 for (j
= 0; j
< n_entries
; j
++)
936 xdg_uint32_t mimetype_offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 8 * j
);
937 xdg_uint32_t parents_offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 8 * j
+ 4);
939 if (strcmp (cache
->buffer
+ mimetype_offset
, mime
) == 0)
941 xdg_uint32_t parent_mime_offset
;
942 xdg_uint32_t n_parents
= GET_UINT32 (cache
->buffer
, parents_offset
);
944 for (k
= 0; k
< n_parents
&& p
< 127; k
++)
946 parent_mime_offset
= GET_UINT32 (cache
->buffer
, parents_offset
+ 4 + 4 * k
);
948 /* Don't add same parent multiple times.
949 * This can happen for instance if the same type is listed in multiple directories
951 for (l
= 0; l
< p
; l
++)
953 if (strcmp (all_parents
[l
], cache
->buffer
+ parent_mime_offset
) == 0)
958 all_parents
[p
++] = cache
->buffer
+ parent_mime_offset
;
965 all_parents
[p
++] = NULL
;
967 result
= (char **) malloc (p
* sizeof (char *));
968 memcpy (result
, all_parents
, p
* sizeof (char *));
974 cache_lookup_icon (const char *mime
, int header
)
977 int i
, min
, max
, mid
, cmp
;
979 for (i
= 0; _caches
[i
]; i
++)
981 XdgMimeCache
*cache
= _caches
[i
];
982 xdg_uint32_t list_offset
= GET_UINT32 (cache
->buffer
, header
);
983 xdg_uint32_t n_entries
= GET_UINT32 (cache
->buffer
, list_offset
);
990 mid
= (min
+ max
) / 2;
992 offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 8 * mid
);
993 ptr
= cache
->buffer
+ offset
;
994 cmp
= strcmp (ptr
, mime
);
1002 offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4 + 8 * mid
+ 4);
1003 return cache
->buffer
+ offset
;
1012 _xdg_mime_cache_get_generic_icon (const char *mime
)
1014 return cache_lookup_icon (mime
, 36);
1018 _xdg_mime_cache_get_icon (const char *mime
)
1020 return cache_lookup_icon (mime
, 32);
1024 dump_glob_node (XdgMimeCache
*cache
,
1025 xdg_uint32_t offset
,
1028 xdg_unichar_t character
;
1029 xdg_uint32_t mime_offset
;
1030 xdg_uint32_t n_children
;
1031 xdg_uint32_t child_offset
;
1034 character
= GET_UINT32 (cache
->buffer
, offset
);
1035 mime_offset
= GET_UINT32 (cache
->buffer
, offset
+ 4);
1036 n_children
= GET_UINT32 (cache
->buffer
, offset
+ 8);
1037 child_offset
= GET_UINT32 (cache
->buffer
, offset
+ 12);
1038 for (i
= 0; i
< depth
; i
++)
1040 printf ("%c", character
);
1042 printf (" - %s", cache
->buffer
+ mime_offset
);
1046 for (i
= 0; i
< n_children
; i
++)
1047 dump_glob_node (cache
, child_offset
+ 20 * i
, depth
+ 1);
1052 _xdg_mime_cache_glob_dump (void)
1055 for (i
= 0; _caches
[i
]; i
++)
1057 XdgMimeCache
*cache
= _caches
[i
];
1058 xdg_uint32_t list_offset
;
1059 xdg_uint32_t n_entries
;
1060 xdg_uint32_t offset
;
1061 list_offset
= GET_UINT32 (cache
->buffer
, 16);
1062 n_entries
= GET_UINT32 (cache
->buffer
, list_offset
);
1063 offset
= GET_UINT32 (cache
->buffer
, list_offset
+ 4);
1064 for (j
= 0; j
< n_entries
; j
++)
1065 dump_glob_node (cache
, offset
+ 20 * j
, 0);