| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
|
|
| |
Using -Bsymbolic-functions to avoid internal PLT indirection is an
easy win. This is the same code that both GLib and GTK+ have.
https://bugzilla.gnome.org/show_bug.cgi?id=689648
|
| |
|
|
|
|
|
|
| |
All tests in that file pass, lets keep it that way.
This is as of Unicode 6.1.0. 6.2 needs more work.
|
| |
|
| |
|
| |
|
|
|
|
| |
Oops!
|
|
|
|
| |
Oops!
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
In UAX#29 revision 19 (reissued for Unicode 6.1.0), Prepend and Extend
GCB properties have been dropped for Thai/Lao. So, drop the hard coding
accordingly.
https://bugzilla.gnome.org/show_bug.cgi?id=576156
|
|
|
|
| |
This was spotted by Coverity.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
Gobject-introspection encounters it without the definitions in the
including code and fails.
|
|
|
|
| |
See bug 688356
|
| |
|
| |
|
|
|
|
| |
autoconf is going to drop support for .in at some point.
|
|
|
|
|
|
| |
Colors that were defined to different values in Pango than they were in
the CSS specification have been updated to conform to the CSS specs.
See http://dev.w3.org/csswg/css3-color/#svg-color for the color values.
|
|
|
|
|
| |
Color names taken from
http://dev.w3.org/csswg/css3-color/#svg-color
|
|
|
|
|
|
|
| |
We can't call pango_xft_shutdown display while holding
the same lock it is taking.
https://bugzilla.gnome.org/show_bug.cgi?id=687470
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
We now need a load_fontset() implementation for the Windows backend as the
pango.aliases handling is moved to pangowin32-fontmap.c since commit
831611158, so that it will go through the pango.aliases file and load
the appropriate fontset needed to display various texts.
This basically copies pango_font_map_load_fontset() from pango-fontmap.c
and adds the old (pre-commit 831611158) pango_font_map_fontset_add_fonts(),
but using pango_win32_font_map_load_font() in place of
pango_font_map_load_font().
|
| |
|
|
|
|
|
|
|
| |
It is just super-annoying that one cannot even do 'make dist'
and get a working tarball. There has to be a better to do this;
for now, just remove this, no other gtk-doc-using module is
doing this.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
After adding pango_shape_full() and passing full paragraph text to
HarfBuzz, we were getting cluster values relative to beginning to
paragraph back from HarfBuzz, and I forgot to account for that. Caused
severe crashes.
Fixes:
Bug 684959 - crash after "Add pango_shape_full()"
and the dupes...
|
| |
|
|
|
|
| |
This reverts commit 150691036aab59044d05ed3fa29785910249bb9a.
|
| |
|
|
|
|
|
|
| |
The doc check script gets tripped up by some struct members
that gtk-doc considers undocumented, and for some reason,
forcing a doc rebuild in make dist does not work either.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Variant of pango_shape() that takes the full paragraph text as input.
This is then passed in entirety to HarfBuzz, which would allow certain
cross-run interactions (in Arabic for example).
When combined with upcoming HarfBuzz 0.9.5+, this fixes:
Red Hat Bug 858736 - [Spanish] Stray dotted circle rendered
https://bugzilla.redhat.com/show_bug.cgi?id=858736
and partially:
Bug 313181 - color changes break arabic shaping
https://bugzilla.gnome.org/show_bug.cgi?id=313181
|
|
|
|
| |
After 88692e4e, every make invocation was rebuilding everything. Fix.
|
|
|
|
| |
For its PangoCoreTextFontMap* member.
|
|
|
|
| |
Patch from galtgendo.
|
|
|
|
| |
This reverts commit 9c1e5d60a078db4b08bd7247004e759ff60cc534.
|
| |
|
|
|
|
|
|
| |
The doc check script gets tripped up by some struct members
that gtk-doc considers undocumented, and for some reason,
forcing a doc rebuild in make dist does not work either.
|
|
|
|
|
|
|
| |
The maps list is protected with a mutex.
registered_engines, dlloaded_engines and dlloaded_modules are
read-only after init_modules, which is protected with g_once_init.
Finally, engine creation is protected with another mutex.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The code was loading and parsing the system config file
every single time, and has not correctly handling mixed
requests for user and system config entries.
This commit reworks the code so that the configuration is
loaded only once, in threadsafe fashion.
pango_config_key_get_system is no longer using the same
hash table, but reloads its data every time - this is
not a really problem, since this function is only used
in pango-querymodules.
|
| |
|
| |
|
|
|
|
|
|
| |
-Include glib/gstdio.h for g_fopen
-Make up for a missed #endif
-correct an internal function name, pango_load_aliases to load_aliases
|
| |
|
| |
|