| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
Now that we're using new-style deprecations, we have to fight
a little harder to get rid of them inside pango.
|
|
|
|
|
|
|
| |
Update the copyright notice of this file, to reflect things more
accurately and to give more proper acknowledgement.
Please see: https://bugzilla.gnome.org/show_bug.cgi?id=767587#c16
|
|
|
|
|
|
|
|
|
|
| |
Add a configure check to see whether compiler directives are available
for exporting symbols, and use them if so. Likewise, update
the Visual Studio projects and config.h.win32.in to do likewise for
Windows builds.
We can then drop the .def files that were used to export symbols on
Windows builds, which should clean up things a bit.
|
|
|
|
|
| |
This is to ensure that we can override the verion macros to export the symbols
later.
|
|
|
|
|
|
|
| |
From the last commit, decorate all the symbols with the version
macros. Also add the version macros in the docs.
https://bugzilla.gnome.org/show_bug.cgi?id=767587
|
|
|
|
|
|
|
|
|
|
|
| |
This header is used for:
-Indicating what stable versions an API is introduced at and after 1.30.0.
-Indicating what APIs are deprecated, and which stable release series
they are deprecated if the deprecation is done after 1.30.0. This replaces
the use of G_DEPRECATED* in Pango.
-Use in a later commit to be used to export symbols directly via compiler
directives instead of doing so over .def files, for compilers that do not
support exporting via regex expressions.
|
| |
|
|
|
|
|
|
| |
http://lists.w3.org/Archives/Public/www-style/2014Jun/0312.html
Thanks Lapo for pointing out this omission!
|
| |
|
|
|
|
|
|
| |
Since Pango coverage is initialized to PANGO_COVERAGE_NONE, it makes no
sense to perform calls to pango_coverage_set for characters that will
be set (or rather left) to none.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
This uses the newly-added autotools module so that the file list
for introspection using Visual Studio is now generated during
'make dist', along with the g-ir-scanner command line to generate
the .gir files and the compilation of the .typelib files.
This would help to simplify maintenance of the introspection
builds in Visual Studio, which is done using NMake.
https://bugzilla.gnome.org/show_bug.cgi?id=764984
|
|
|
|
|
|
|
|
| |
Add finalizers to PangoWin32FontMap
Patch from Geoffrey Yerem.
Part of https://bugzilla.gnome.org/show_bug.cgi?id=764089
|
|
|
|
|
| |
Unicode says: "U+1680 ogham space mark is unusual in that it is generally
rendered with a visible horizontal line, rather than being blank."
|
|
|
|
|
| |
It was referring to itself for further description, but I assume the
intent was to refer to pango_layout_line_x_to_index().
|
| |
|
|
|
|
| |
This brings us back in sync with GUnicodeScript.
|
|
|
|
| |
No changes here.
|
| |
|
|
|
|
|
| |
Avoid determining the Unicode type repeatedly. Instead,
just get the type once and filter out the classes we want.
|
|
|
|
|
| |
We already have the Unicode type, so we can determine
isspace without doing the full type determination again.
|
|
|
|
| |
Instead of going through a wrapper.
|
|
|
|
|
| |
Some compilers insist on using a number for a guint16 return type, and does
not like using NULL in place of 0, so fix this.
|
|
|
|
|
|
| |
Harfbuzz 1.0.2 had errors in this area that gave us crashes
in the testsuite (testiter). Therefore, only use this feature with
Harfbuzz 1.0.3 or newer.
|
|
|
|
|
|
| |
"Fixes" https://bugzilla.gnome.org/show_bug.cgi?id=541608
Real fix went to HarfBuzz.
|
|
|
|
| |
Part of https://bugzilla.gnome.org/show_bug.cgi?id=541608
|
|
|
|
|
|
| |
Much better now.
Now I see issues in justify_words() again :(.
|
| |
|
|
|
|
| |
Was totally b0rked. Much better now. Should be good enough...
|
| |
|
|
|
|
|
|
|
| |
Bug 753772 - Justify option inserts spaces in wrong positions
https://bugzilla.gnome.org/show_bug.cgi?id=753772
Went unnoticed for 8 years...
|
|
|
|
| |
Ouch!
|
| |
|
|
|
|
|
|
|
| |
Libthai is not threadsafe (thought it probably can easily be made to be).
As such, add mutex around th_brk() calls.
test-pangocairo-threads passes again.
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=753167
|
|
|
|
|
|
|
| |
So that we can set attributes (e.g. colors) to the marks and their bases
indecently.
https://bugzilla.gnome.org/show_bug.cgi?id=541608
|
|
|
|
|
| |
Instead of relying on the fact that marks have the same cluster number
as their bases, which will change in the next commit.
|
| |
|
|
|
|
|
|
|
| |
For foreground and background colors, allow #rrggbbaa to specify
a color with alpha. This will be returned as two attributes, for
color and alpha. underline_color and strikethrough_color are still
limited to #rrggbb, with no alpha component.
|
|
|
|
|
|
| |
Support alpha and background_alpha attributes on span elements.
The value can either be a plain integer between 1 and 65536 or
a percentage like 50%.
|
|
|
|
|
|
| |
Make PangoCairoRenderer use cairo_set_source_rgba() to apply
colors with alpha. Note that this means we are only using the
alpha if the corresponding color has been set.
|
|
|
|
|
|
| |
Add pango_renderer_get/set_alpha to allow associating alpha
values with the various colors. The alpha is stored as a value
between 0 and 65536, with 0 meaning 'unset'.
|
|
|
|
| |
Add attributes types for foreground and background alpha.
|
| |
|
|
|
|
|
|
| |
It was combining the last two clusters. Ouch!
Test added already. Passes now.
|
| |
|
|
|
|
|
| |
HarfBuzz buffer comes with a properly set up Unicode funcs
these days. So, trust it.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=738505
Patch from Matthias Clasen, based on early patch from Akira TAGOH.
There's room for improvement in how this is done, but it works now
for simple cases, which is what most people will be using it for.
Finally!
|