| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Patch from Rafał Mużyło.
|
|
|
|
|
|
|
|
| |
When handling PANGO_ATTR_SCALE on a font description with an absolute
size, properly preserve the absolute flag when setting the scaled
size.
https://bugzilla.gnome.org/show_bug.cgi?id=708778
|
|
|
|
| |
PANGO_GRAVITY_NORTH instead of PANGO_GRAVITY_SOUTH?
|
|
|
|
|
|
| |
The annotation for pango_layout_set_attributes should be
'transfer none', not 'transfer full'. Pointed out in
https://bugzilla.gnome.org/show_bug.cgi?id=707874
|
|
|
|
|
|
|
|
|
|
|
| |
layouts get initialized with text == NULL as an optimization (avoid a
malloc). But pango_layout_set_text (layout, NULL, 0); will set the text
to "", so it is impossible to set a NULL text.
Fxies crashers in various places that assume NULL return values never
happen.
https://bugzilla.gnome.org/show_bug.cgi?id=707659
|
|
|
|
|
|
|
|
|
| |
Since pango_scan_int() would return FALSE if the scanned value was over
INT_MAX; and the reference out value's type is an int, (n > INT_MAX)
test is unecessary.
If pango_scan_int() behavior was to be changed to accept long
int, an explicit test should be added (or the PangoAttrSize
type updated to handle long int too).
|
|
|
|
|
| |
Since the value is stored as an int, no reason to use any other value
than INT_MAX.
|
|
|
|
| |
Patch from Akira TAGOH.
|
|
|
|
| |
Fix all-RTL reordering. Oops!
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Back in the days we used to always use the font from surrounding runs
for any space characters. I changed that when I found that rendering
MathML results in hexboxes when a font doesn't have the space characters
needed...
Forward to these days: HarfBuzz will silently compatibility-decompose
any space character to the ASCII space. And looks like it's important
for Mongolian to not break runs around NNBSP. As such, go back to never
choosing a font just for space characters.
Seems to work, though more testing is needed.
Bug 701652 - wrong shaped classic Mongolian word suffixes
|
| |
|
|
|
|
|
|
|
| |
Newer Unicode specs (6.2 for instance)
say that one should not break between letter or numbers
that are adjacent. This also makes us to match word selection
with many text capable apps.
|
|
|
|
|
|
|
|
| |
This partially reverts 25ad29b488d3e2ae10e4fee25602409fa39207f1
which itself was a revert. Apparently we want the change in
CoreText, but not on Win32...
See mailing list discussion.
|
| |
|
| |
|
|
|
|
| |
Based on patch from Chun-wei Fan.
|
| |
|
| |
|
|
|
|
| |
Patch from Chun-wei Fan.
|
|
|
|
| |
libpang-1.0.la in _gir_LIBS
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit b359bb484bc7d27a835cfd6d9ffcaed582ebf2ae.
Bug 694918 - Pango layout is rendered in wrong size if we have
non-identity cairo transformation matrix
I was wrong blindly copying this logic from pangocairo-fc into
pangocairo-win32 and pangocairo-quartz.
The reason we need the it in pangofc is because the fontsize we
get back from fontconfig is multiplied by the ctm scale factor,
so we undo it when loading the font. The same is NOT true about
the non-fc backends, so the copying was totally wrong.
|
| |
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=694626
|
|
|
|
|
|
|
| |
This is needed as a base class sometimes needs to invalidate
the fontmap.
https://bugzilla.gnome.org/show_bug.cgi?id=694626
|
|
|
|
|
|
| |
When win32 fontmap goes away, font->fontmap becomes NULL. Deal
with that. Technically speaking, we need to port this to GWeakRef
like the other fontmaps do. But this would do for now.
|
|
|
|
|
|
|
| |
.def files are mostly used on Win32. Whitelist the win32 font map
function in the check.
Bug 694095 - Fails make check: -pango_cairo_win32_font_map_get_type
|
|
|
|
| |
These have now been put into Harfbuzz upstream
|
|
|
|
| |
This is needed to build the introspection files for PangoCairo on Windows.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit 2dc0c3dbb1c389c3a3ba12a5c5c85f21dca46e84.
This doesn't make sense. It ends up in infinite recursion because
pangofc is calling lock_face on pangoft2, which would recursively
call it on pangofc...
Was causing hang at startup of Inkscape.
If the deprecated warnings are an issue, they need to be dealt with
in some other way.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Like this:
<match target="font">
<edit name="pangofontfeatures" mode="append">
<string>smcp</string>
<string>ss20</string>
</edit>
</match>
Finally we are starting to see new features coming out of the HarfBuzz integration...
|
|
|
|
| |
Replace it with PANGO_GET_UNKNOWN_GLYPH.
|
|
|
|
| |
Replace it with pango_fc_font_lock/unlock_face calls.
|
| |
|
|
|
|
| |
gcc doesn't like #ifndef LANGUAGE(x, y, z).
|
|
|
|
|
| |
Add the new paragraph_text/length parameters to the script_shape
implementation in the fallback engine.
|
|
|
|
|
| |
g_type_init has been deprecated in GLib 2.35.x. Use a
version check here to avoid depending on unstable GLib.
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
By way of declaring fontmaps NOT threadsafe, and making
pango_cairo_font_map_get_default() return a thread-private fontmap.
test-pangocairo-threads doesn't crash anymore, if used with fontconfig
master.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
memory"
This reverts commit 7ed3cb89923c376d8b30ae3f977ab9e1a231e430, after
resolving conflicts.
The patch introduced some lifecycle management issues that could easily
cause crashing. Reverting till the issues are resolved.
|
| |
|
|
|
|
|
|
|
| |
If we are in fallback mode, with a font that has to space glyph, then
looking up engine/font for any character not in the font was failing
badly. In that case, if there's only one engine (which is the case
these days), just choose it.
|
|
|
|
|
| |
After commit 7c1d6208a4b42ef3a3cce0d8dffb214c7a321e5e, the first matched
font was being prepared twice, causing misconfiguration.
|
|
|
|
|
| |
This reverts commit 7274bd56db3ce2f02594687a307d22aa99c787c4.
Whatever the problem was with g-ir-scanner, it's resolved.
|
|
|
|
| |
Refcount cached patterns.
|
|
|
|
| |
Fix OS X relocation.
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=679299
|