With the removal of the in-directory cache files, and the addition of
per-user cache directories, there is no longer any reason to preserve the
giant global cache file. Eliminating of this unifies the cache structure
and simplifies the overall caching strategies greatly.
Normalized directory names offer protection against looped directory trees
but cost enormous numbers of system calls (stat per file in the hierarchy).
Also, cache file directory name contents are validated each time the
directory is modified, don't re-validate every time the cache file is loaded
with an access and stat call.
Check if pattern is not null before using it (Coverity defect #1883).
Fix memory leak with hash collision (Coverity defect #1829).
Fix memory leak when bail cases (Coverity defect #1828).
Don't leak directory name (Coverity defect #1827).
reviewed by: plam
that aren't hardcoded into fontconfig, but generated by fontconfig
clients: keep another array of user-defined objects (indexed after the
built-in objects).
Fix compilation warning (uninitialized variable).
Add comment.
data format. Also match subdirectories when consuming cache
information. Also check dates for global cache: a dir is out of date if
it is newer than the global cache; scan it manually if that's the case.
'object' table (strings pointed to by FcPatternElt->object and used as
keys) and loading of object table from cache file if more strings are
present in cache file than in current version of fontconfig. Hash the
object table in memory.
probably all right to use the global cache as it was previously and
just store filenames and font info, as long as no mmap cache exists in
the directory. Of course, if an mmap cache exists, use that instead.
If a directory cache does not exist or is invalid, load the fonts for just
that directory using the old codepath.
Fix premature free of the FcPatterns belonging to the FcFontSet which we
create from the mmapped files.
and distribute bytes for each directory from a single malloc for that
directory. Store pointers as differences between the data pointed to
and the pointer's address (s_off = s - v). Don't serialize data
structures that never actually get serialized. Separate strings used
for keys from strings used for values (in FcPatternElt and FcValue,
respectively). Bump FC_CACHE_VERSION to 2.
cache. Add *Read and *Write procedures which mmap in and write out the
fontconfig data structures to disk. Currently, create cache in /tmp,
with different sections for each architecture (as returned by uname's
.machine field. Run the fc-cache binary to create a new cache file;
fontconfig then uses this cache file on subsequent runs, saving lots of
memory. Also fixes a few bugs and leaks.
ids can be positive (for static strings) or negative (for dynamic
strings). Static strings belong to a single buffer, while dynamic
strings are independently allocated.
This patch allows the fundamental fontconfig data structures to be
serialized. I've converted everything from FcPattern down to be able to
use *Ptr objects, which can be either static or dynamic (using a union
which either contains a pointer or an index) and replaced storage of
pointers in the heap with the appropriate *Ptr object. I then changed
all writes of pointers to the heap with a *CreateDynamic call, which
creates a dynamic Ptr object pointing to the same object as before.
This way, the fundamental fontconfig semantics should be unchanged; I
did not have to change external signatures this way, although I did
change some internal signatures. When given a *Ptr object, just run *U
to get back to a normal pointer; it gives the right answer regardless
of whether we're using static or dynamic storage.
I've also implemented a Fc*Serialize call. Calling FcFontSetSerialize
converts the dynamic FcFontSets contained in the config object to
static FcFontSets and also converts its dependencies (e.g. everything
you'd need to write to disk) to static objects. Note that you have to
call Fc*PrepareSerialize first; this call will count the number of
objects that actually needs to be allocated, so that we can avoid
realloc. The Fc*Serialize calls then check the static pointers for
nullness, and allocate the buffers if necessary. I've tested the
execution of fc-list and fc-match after Fc*Serialize and they appear to
work the same way.
Add detection of iconv
Document new selectfont elements
Switch to UTF-8 in comment
Add fullname, and family/style/fullname language entries
Respect selectfont/*/glob
Add support for selectfont
Add multi-lingual family/style/fullname support
Expose FcListPatternMatchAny (which selectfont/*/pattern uses)
Add new FcPatternRemove/FcPatternAppend. FcObjectStaticName stores computed
pattern element names which are required to be static.
However FcConfigUptoDate() doesn't seem to work. See the attached patch.
First there's an obvious misplaced parenthesis making it return always
false, and second, even this call fails to detect font changes (e.g.
adding a new font to /usr/X11R6/lib/X11/fonts/truetype). The patch
should fix that as well. The problem seems to be triggered by my
fonts.conf specifying only /usr/X11R6/lib/X11/fonts , and therefore
config->configDirs doesn't include subdirs, unlike config->fontDirs.
Oops. Left 'newest.set' unset, which would miscompute the newest file
Add FcGetPixelSize to extract correct pixel size from bdf/pcf font
properties (which report the wrong value in current FreeType)
Don't attempt to check for empty glyphs in non-scalable fonts; they have no
outlines...
when trying to save the cache if config->cache is NULL, which happens
if FcConfigHome() is NULL. Guard against that by using the temp folder
in that case.
LISTING requires that the font Contain all of the pattern values, where
Contain is redefined for strings to mean precise matching (so that
Courier 10 Pitch doesn't list Courier fonts)
"Contains" for lang means both langs have the same language and either the
same country or one is missing the country
On Windows with gcc (a.k.a. mingw) build as a DLL.
We don't want to hardcode the fonts.conf file location in the DLL, so we
look up the DLL location at run-time in a DllMain() function. The
fonts.conf location is deduced from that.
The colon can't be used as path separator on Windows, semicolon is used
instead. File path components can be separated with either slash or
backslash. Absolute paths can also begin with a drive letter.
Add internal function FcStrLastSlash that strrchr's the last slash, or
backslash on Windows.
There is no link() on Windows. For atomicity checks, mkdir a lock directory
instead.
In addition to HOME, also look for USERPROFILE.
Recognize the special font directory token WINDOWSFONTDIR, to use the
system's font directory.
Remove the fontconfig-def.cpp that was obsolete. Add fontconfig.def(.in),
without internal functions.
Add a fontconfig-zip(.in) script, used to build a binary distribution.
Keith Packard
Patrick Lam
Keith Packard
Tor Lillqvist
Manish Singh
Carl Worth
James Su
David Dawes