f11a184104
Don't kill fonts because they don't declare an element that's being matched
...
on.
reviewed by: plam
2006-02-10 19:40:11 +00:00
879af7060b
Define and pass O_BINARY to open if appropriate, for those platforms that
...
need it. Reported by Doodle.
2006-02-09 18:44:14 +00:00
c7490074c5
Fix attempt to close -1. Don't unboundedly grow bad caches (reported by
...
fcrozat).
2006-02-09 16:19:42 +00:00
a8c425301a
Fix problem with missing 'en' due to euro.patch: change cache file format
...
slightly to coincide with that generated by fc-lang.
2006-02-09 15:25:57 +00:00
b10e77628c
Fix warning.
2006-02-08 03:34:17 +00:00
5c3deb2985
Don't reject dirs that can't be normalized (fixes lilypond, and is correct
...
in the context of application font directories.)
Use normalized directory name exclusively in FcCacheReadDirs.
reviewed by: plam
2006-02-07 22:09:01 +00:00
efb11b36c4
Perf optimizations. Inline FcValueCanonicalize, reduce FcValueListPtrU
...
usage, remove redundant cast.
reviewed by: plam
2006-02-07 21:15:33 +00:00
6cc02fe6b9
src/fccharset.c (FcLangCharSetPopulate, FcCharSetInsertLeaf)
...
Fix missing FcCacheBankToIndex in FcCharSetInsertLeaf. Declare extern for
static arrays as arrays, not pointers. (Part of the fix for 'fonts
don't have en' issue after Euro patch.)
(I forgot to commit the ChangeLog last time.)
reviewed by: plam
2006-02-07 20:56:48 +00:00
a81f23c0ce
Fix hidden variable warning.
2006-02-07 03:53:32 +00:00
799157dbbf
Remove de-escaping logic because FcCacheWriteString doesn't escape anyway.
...
Do blockwise reading instead of byte-wise for performance.
2006-02-07 02:33:57 +00:00
8b413bb62c
Takashi Iwai <tiwai@suse.de>
...
Don't loop infinitely on recursive symlinks (client-side).
2006-02-07 02:22:50 +00:00
660acf8f22
Don't loop infinitely on recursive symlinks.
...
reviewed by: plam
2006-02-06 23:11:41 +00:00
fff5a5af30
Skip subdirs when skipping over stale bits of global cache. Introduce state
...
machine into FcGlobalCacheDir to avoid doing inappropriate operations
on global dir entries, e.g. writing out an out-of-date cache entry.
reviewed by: plam
2006-02-06 22:44:02 +00:00
98592bbb1d
Hoist FcFileIsDir check out of FcFileScanConfig loop.
...
reviewed by: plam
2006-02-06 21:52:15 +00:00
a0aa54f6ee
Don't rescan when trying to normalize a non-declared font dir. Don't add
...
font dirs multiple times (even if they're aliased).
reviewed by: plam
2006-02-06 19:25:45 +00:00
86e75dfb5d
Explain apples/oranges comparison and fix compilation error.
...
reviewed by: plam
2006-02-06 14:44:46 +00:00
f076169d19
Insert check for integer overflow in # of fonts.
...
reviewed by: plam
2006-02-06 14:14:21 +00:00
788c4af232
Make 'make distcheck' work with automake 1.6.3.
...
reviewed by: plam
2006-02-05 04:11:08 +00:00
68355f3877
src/fccache.c (FcGlobalCacheLoad, FcGlobalCacheSave, FcDirCacheConsume,
...
FcDirCacheWrite)
Check I/O call return values and eliminate unused variable warnings.
reviewed by: plam
2006-02-05 02:57:21 +00:00
c4c47a7654
src/fccfg.c (FcConfigAppFontAddFile, FcConfigAppFontAddDir)
...
Fix memory leak.
reviewed by: plam
2006-02-04 00:09:42 +00:00
a8e4d9eb39
Gracefully handle the case where a cache asserts that it has a negative
...
number of fonts, causing overflow.
reviewed by: plam
2006-02-04 00:04:00 +00:00
1af0f5741a
Fix double free in error case.
2006-02-03 23:47:37 +00:00
80ba0571f3
Stephan Kulow <coolo@suse.de> reviewed by: plam
...
Replace 'stamp' target with mkinstalldirs.
2006-01-31 07:30:23 +00:00
28aefd013d
Toast broken global cache files.
...
reviewed by: plam
2006-01-31 07:16:22 +00:00
3616cbe251
Actually use the directories that I added to the global cache. Fixes 'make
...
check'.
2006-01-31 04:42:20 +00:00
275cf6cd83
Don't stop scanning if a directory in fonts.conf doesn't exist, because
...
subsequent directories might exist.
reviewed by: plam
2006-01-31 04:07:45 +00:00
81d97094cf
Make global cache work again after putting dir names into global cache (I
...
misplaced the recording of a file offset).
2006-01-30 16:31:56 +00:00
2c4e012497
Update fc-cat to handle subdir lists in global cache file.
...
Another FcCacheReadString return value check.
2006-01-30 15:59:17 +00:00
c5411c4cae
Make fccache more resilient to broken cache files by checking return value
...
of FcCacheReadString all the time.
reviewed by: plam
2006-01-30 15:44:13 +00:00
946478e1a7
Remove references to dead fontconfig(3) manpages in other fontconfig
...
manpages.
reviewed by: plam
2006-01-30 14:43:04 +00:00
af7a965f94
Fix world's tiniest typo in code example.
...
reviewed by: plam
2006-01-30 04:51:22 +00:00
30c4189deb
Fix global cache reads of subdirectories.
...
reviewed by: plam
2006-01-30 04:47:17 +00:00
971cf18018
Add documentation for FcConfigNormalizeFontDir.
...
Write directory information to global caches to fix make check (reported by
Ronny V. Vindenes). This changes the global cache format again.
2006-01-30 04:27:53 +00:00
97293e07dd
Move FcConfigNormalizeFontDir call so that it doesn't result in infinite
...
recursion (reported by Ronny V. Vindenes).
2006-01-27 05:47:59 +00:00
3cf9f5cec3
Add a couple of missing normalizations to make fc-cache work right; only
...
scan subdirectories once.
2006-01-27 00:27:37 +00:00
4073203deb
Don't crash on non-existant directories in global cache.
...
reviewed by: plam
2006-01-26 16:11:41 +00:00
f468f568b4
Stop trampling the directory name when writing out caches. (with Mike
...
Fabian:) Beef up FcConfigNormalizeFontDir to scan subdirs when
necessary. Don't scan directories that can't be normalized.
2006-01-26 16:09:12 +00:00
575a37b797
Fix additional memory leaks reported by Ronny V. Vindenes: don't invoke
...
FcValueSave on hashed static strings in FcPatternAddWithBinding.
Add another st_dev check in FcDirCacheOpen.
2006-01-25 14:52:49 +00:00
16a71eff3e
Treat zh-hk fonts differently from zh-tw fonts. This patch may cause
...
fontconfig to treat A-X fonts differently from A-Y fonts; please mail
the fontconfig list if this causes any problems.
reviewed by: plam
2006-01-25 02:54:37 +00:00
6f9fcb5186
Fix memory leaks reported by Ronny V. Vindenes.
2006-01-25 02:33:46 +00:00
986e35979e
Fix for unaligned memory accesses.
...
reviewed by: plam
2006-01-19 19:20:30 +00:00
58bdd29619
Properly order the FcConfigAddFontDir and FcConfigNormalizeFontDir calls to
...
avoid crashes.
reviewed by: plam
2006-01-18 15:41:15 +00:00
eadadf489a
Fix segfault when consuming zero-length caches in fc-cat (which has no
...
config, so FcConfigAddFontDir shouldn't be called.)
2006-01-15 05:31:58 +00:00
8a0b0ed6d0
Compare device numbers as well as inodes. Always normalize directory names
...
before comparing them.
Allocate extra space for appended '/' in directory name.
reviewed by: plam
2006-01-14 21:23:03 +00:00
df3efc11a9
Explicitly add font dirs to config.fontDirs even if they're empty. Set
...
current config in fc-cache.c. Fix treatment of cache directory as read
from cache file; don't use string equality to determine if we have the
right file, use inode equality.
2006-01-10 13:15:05 +00:00
cd9bca6970
Normalize font dirs by using the form, as given in fonts.conf, and recorded
...
in FcConfig's fontDirs string set, as canonical.
Actually update config.fontDirs as font directories are scanned.
2006-01-09 13:58:04 +00:00
5576a5873d
Fix matching bug when multiple elements match; don't use the sum of all
...
scores, but the best score.
Also more perf opts, e.g. evaluate best over all font sets rather than on a
per-set basis (and other changes).
2006-01-08 10:58:30 +00:00
a6d3757d98
Fix matching bug when multiple elements match; don't use the sum of all
...
scores, but the best score.
Also more perf opts, e.g. evaluate best over all font sets rather than on a
per-set basis (and other changes).
2006-01-08 10:58:05 +00:00
05a98eaf4b
Properly skip past dir caches that contain zero fonts (as occurs in global
...
caches.) Reported by Mike Fabian.
2006-01-08 10:50:51 +00:00
5fe09702f4
Print out full pathname in fc-match -v as well. Reported by Frederic
...
Crozat.
Fix bug where fc-match crashes when given __DUMMY__ property to match on.
(I added the __DUMMY__ string to enable callers of FcObjectToPtrLookup to
distinguish an error return from a successful return. -PL)
reviewed by: plam
2006-01-07 06:36:24 +00:00
Patrick Lam