| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
Shell globs in install-commands are not being expanded which is causing
generation of the pango.modules file to fail.
|
| |
|
|
|
|
|
|
| |
Same fix as needed for gdk-pixbuf; the module cache was actually empty
because pango-querymodules needs a library only available inside DESTDIR
at the time we want to run it.
|
|
|
|
|
| |
This should be done automatically, but because we use DESTDIR it does not
get executed.
|
|
|
|
|
|
| |
Commit message written by Colin Walters <walters@verbum.org>
https://bugzilla.gnome.org/show_bug.cgi?id=667960
|
| |
|
|
|
|
|
| |
Oversight in my fallback fix, this resulted in things being wrongly
destroyed due to unrefs. Oops.
|
|
|
|
| |
This was forgotten in commit aecbe27c1b08f517c0e05f03308d3ac55cef490c
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
The fallback failed when a "small caps" font was requested. This commit
improves the fallback support. When the first fallback, trying Sans with
the same style fails, we reset the variant, weight and stretch to
default values and try again. With Sans we should always be able to
adhere to the requested style.
Last but not least, output a sensible error message if all fallbacks
fail instead of simply crashing on a NULL pointer somewhere.
|
|
|
|
|
|
|
|
| |
The documentation says 'This array should be freed with g_free() when
it is no longer needed', which really translates to (transfer
container) in introspection annotation jargon.
https://bugzilla.gnome.org/show_bug.cgi?id=677407
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=677407
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=677407
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=677407
|
|
|
|
|
|
|
|
| |
Change the X fontmap code to look for the aliases file
in $XDG_CONFIG_HOME/pango/pangox.aliases instead of
~/.pangox_aliases
https://bugzilla.gnome.org/show_bug.cgi?id=675400
|
|
|
|
|
|
|
| |
Look for the rcfile in $XDG_CONFIG_HOME/pango/pangorc
instead of ~/.pangorc.
https://bugzilla.gnome.org/show_bug.cgi?id=675400
|
|
|
|
|
| |
Same fallback as already used in CoreText font descriptor to Pango
font descriptor conversion.
|
|
|
|
|
|
|
|
|
|
| |
When run as a system trigger, we really don't want to be consulting
e.g. /root/.pangorc. Also, the OSTree build system can run as
non-root, but we don't in this case want to look at /home/user since
the user may not even exist in the password database inside the
chroot.
https://bugzilla.gnome.org/show_bug.cgi?id=667960
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=672133
|
|
|
|
|
|
|
| |
Setting -I$(top_builddir) before -I$(top_srcdir) causes that
g-ir-scanner picks up boxing definitions generated by glib-mkenums.
https://bugzilla.gnome.org/show_bug.cgi?id=672133
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
Part of Bug 454017 - Make relative paths in pango.modules and pangorc
files work relative to the file
Patch from Anselm Kruis.
|
| |
|
|
|
|
|
|
|
|
|
| |
This commit makes the CoreText shaping engine much more robust against
the output from the CoreText typesetter. Refer to the source code
comments for exhaustive details.
Furthermore, this patch fixing cursor navigation in strings with
zero-width spaces.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Patch written in collaboration with Michael Natterer.
Some CTFontDescriptors apparently do not have a style name or font
family name set. This patch makes the code take such corner cases into
account. The font family problem only appears to occur on Snow Leopard
systems, we try to fall back on the font name (postscript name), if that
fails, we fall back on a default fallback.
In some cases a coverage is also not set. (This seems to happen when we
cannot map a broken font back to a correct font descriptor). In such
cases we simply return an empty PangoCoverage, which will likely cause
the engine to fallback on a different font.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We implement font fallbacks by making use of
CTFontCopyDefaultCascadeList, a public, non-exposed, symbol in the
CoreText library. We might want to handle this differently.
The code now also uses a PangoCoreTextFontsetKey and
PangoCoreTextFontKey, similar to the font config backend. This has also
helped to clean up the code in general.
This patch contains two fixes by Michael Natterer:
1. Fix a bug in pango_core_text_font_key_equal() where the return value
of memcmp() was wrongly interpreted.
2. Cache PangoCoverage in PangoCoreTextFont.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
This is fixed by using the characters obtained from CoreText's mapping from
glyphs back to the original string instead of blindly iterating over the
original gchar string. Take notice that we have to convert these indices
from the mapping back to a byte index in the original UTF8 string.
Zero-width spaces are already removed from the list of CGGlyphs, so we were
mismatching CGGlyphs and characters from the original string. Using
CoreText's list of string indices instead eliminates the mismatches.
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=667799
|
|
|
|
| |
Make the config entries more like the standard config.h.in
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
When gtk-doc regenerated the pango.types file, it includes
get_type functions for platform-specific things that will
cause link failures.
|
|
|
|
|
| |
Users of pango are not generally expected to use gmodule API,
so don't force them all to link against it.
|
|
|
|
|
|
|
|
|
|
| |
It turns out that getting this value as "Int" is incorrect and resulted
in font traits not being returned in some cases. Without traits, an
italic trait is not set, which caused synthetic oblique fonts to be
created when not necessary.
Also use CTFontSymbolicTraits type in the PangoCoreTextFace structure
and do a stricter bit mask check for certainty.
|
| |
|
|
|
|
|
| |
Say farewell to G_DISABLE_DEPRECATED and the build breakages that it causes.
We now have warnings when we make use of deprecated GLib items
|
|
|
|
|
|
| |
Now that we get compiler warnings for deprecations, we don't
need to use the G_DISABLE_DEPRECATED hammer anymore.
Bye, bye, build breakage.
|
|
|
|
|
|
|
|
| |
This is using the newly introduce G_DEPRECATED macros
from GLib. Traditional deprecation guards are still in
place.
https://bugzilla.gnome.org/show_bug.cgi?id=661612
|
|
|
|
| |
1.29.4 got released off a side-branch, but bump version here too.
|
|
|
|
|
|
|
|
| |
This project will require the use of stdint.h, which is not shipped with
VS2008 (but is shipped with VS2010), so define HAVE_STDINT_H for all
settings in this project so that the stdint.h bundled in
$(srcroot)/build/win32/vs9 will be used when HarfBuzz and FontConfig is
being used.
|
|
|
|
|
| |
Update the info about the Visual Studio projects as VS2010 projects have
landed into master.
|
|
|
|
|
|
|
|
| |
G_UNICODE_COMBINING_MARK was renamed as G_UNICODE_SPACING_MARK soon after
the GLib 2.29.14 release, so fix that in break.c and pango-ot-info.c.
To support GLib on and before 2.29.14, a convenience definition for
G_UNICODE_SPACING_MARK is defined.
|
|
|
|
|
| |
g_atomic_int_exchange_and_add was deprecated in the GLib gatomic.c rewrite,
for GLib 2.29.6, so use g_atomic_int_add for GLib versions 2.29.5+.
|
|
|
|
|
|
|
| |
-Seperate intermediate folders for each project to remove redundant VS
warnings/errors
-Don't create a folder for pango.aliases if the folder already exists
-Don't create pango.aliases if it already exists
|
|
|
|
|
|
|
| |
-This file should have Windows EOL
-Tell people about GNOME Live! page for steps to building Pango with MSVC
(and the GTK+ stack)
-Fix some errors in some of the words used here.
|