| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
| |
|
| |
|
|\
| |
| |
| |
| | |
Various introspection fixes
See merge request GNOME/pango!117
|
| |
| |
| |
| |
| | |
Since Harfbuzz does not generate introspection data, we cannot use it
for our introspected ABI.
|
| |
| |
| |
| |
| | |
These symbols are generating warnings because of missing type
information, but since they are deprecated we can just skip them.
|
| |
| |
| |
| |
| |
| |
| |
| | |
Makes it usable by language bindings, and avoids warnings when
generating introspection data.
We already have the necessary public API, we're just missing a GType for
it.
|
| |
| |
| |
| |
| | |
Makes it usable by language bindings, and avoids a warning when
introspecting the source code.
|
| |
| |
| |
| |
| |
| | |
The `pango_coverage_get_type()` symbol is already public, but declared
in a private header. Unless we move it to a public header the
introspection scanner can't know that the C type maps to a GType.
|
| |
| |
| |
| | |
Otherwise the introspection scanner will be majorly confused.
|
| |
| |
| |
| |
| |
| | |
The unprefixed `script_*` symbols are marked as `skip`, but the
introspection scanner will still warn about them because the annotations
are resolved after parsing the C code.
|
| | |
|
|/ |
|
|
|
|
|
| |
This is problematic when building pango as subproject
of gtk on the win32 ci.
|
|\
| |
| |
| |
| | |
Compiler warnings
See merge request GNOME/pango!116
|
| |
| |
| |
| |
| |
| | |
This reverts commit 315e3d4f6182e4cefde3538ede222e53c000f3e8.
This made pango-view segfault in win32 ci.
|
| |
| |
| |
| |
| |
| | |
Add warning flags for the just fixed warnings
to the build. This is in order to ensure that
pango builds in ci as a subproject of gtk.
|
| |
| |
| |
| | |
Unused-but-set-variable.
|
| |
| |
| |
| | |
This was pretty clearly unintentional.
|
|/
|
|
| |
This triggers compiler warnings.
|
|\
| |
| |
| |
| | |
shape: Simplify font function
See merge request GNOME/pango!115
|
| |
| |
| |
| | |
Instead of setting variation selector to 0.
|
|/
|
|
|
| |
If a function is not set, HarfBuzz will default to the parent’s, so no
need to set functions we don’t customize.
|
|\
| |
| |
| |
| | |
Missing hex boxes
See merge request GNOME/pango!114
|
| |
| |
| |
| |
| |
| | |
If we don't use our own font funcs, hex boxes
don't work, since harfbuzz just replaces missing
glyphs with 0.
|
| |
| |
| |
| |
| |
| | |
It is misleading to see uninitialized pointers
in gdb, so set them to NULL, even if they are
not used.
|
|/
|
|
|
| |
While of no immediate consequence, it is the
right thing to do.
|
|\
| |
| |
| |
| |
| |
| | |
Keep deprecated pango_fc apis in pango-fc-font.h
Closes #394
See merge request GNOME/pango!113
|
| |
| |
| |
| |
| |
| | |
Turns out these are in use.
Closes: https://gitlab.gnome.org/GNOME/pango/issues/394
|
|\ \
| |/
|/|
| |
| | |
coretext: Make “system-ui” alias work
See merge request GNOME/pango!112
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Alias names need to be inserted into the font map or otherwise will be
ignored and mapped to Sans (in pango_core_text_font_map_load_fontset()).
Furthermore, there is no font named ”San Francisco” on macOS, the actual
UI font is a hidden font called ”.SF NS Text” or something like that,
and there an “.AppleSystemUIFont” alias that should be used instead.
|
|/
|
|
|
|
|
| |
Add commandline options for antialiasing,
subpixel order and metrics hinting, so we
can reproduce font rendering issues with
pango-view.
|
|\
| |
| |
| |
| | |
Set design coords for named instances
See merge request GNOME/pango!110
|
|/
|
|
|
|
|
|
|
| |
Fontconfig has FC_INDEX, and harfbuzz has
hb_ot_var_named_instance_get_design_coords.
We can combine the two to set coords for
variable fonts, while taking overridden
variations from both the fontconfig configuration
and the font description into account.
|
|\
| |
| |
| |
| |
| |
| | |
'master'
pango-list: Call setlocale in main function
See merge request GNOME/pango!109
|
|/
|
|
|
|
| |
If it isn't called, GLib will try to convert non-ASCII characters to
ASCII because LC_CTYPE locale is C. Therefore, all non-ASCII characters
are displayed as question marks because the conversion always fails.
|
|\
| |
| |
| |
| |
| |
| | |
Turn off ligatures when letterspacing
Closes #182
See merge request GNOME/pango!108
|
|/
|
|
|
|
|
|
|
| |
When letterspacing, we don't want the letters making
up ligatures to 'stick together', so turn off OpenType
features that request optional ligatures (liga, clig,
dlig).
Closes: https://gitlab.gnome.org/GNOME/pango/issues/182
|
| |
|
| |
|
|\
| |
| |
| |
| | |
Add freetype2 wrap file
See merge request GNOME/pango!106
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The freetype2.wrap file was downloaded using
"meson wrap install freetype2". This adds a wrap
dependency that had already been declared in
meson.build as a fallback for freetype2.
Signed-off-by: Nicholas Harvey <727-FrostyPenguin@users.noreply.gitlab.gnome.org>
|
|\ \
| |/
|/|
| |
| | |
Take font matrix into account
See merge request GNOME/pango!107
|
|/
|
|
|
|
|
|
|
|
|
|
|
| |
When setting up the scale of the hb_font_t, we need
to take both the pango ctm and the font matrix into
account - this used to come for free when we were
calling into cairo for getting glyph metrics. Now,
we use harfbuzz for glyph metrics, so we need to
give it the proper scale info.
The symptom of this was Emoji getting an enormous
width. Curiously, cairo would still render them
at the expected size.
|
|
|
|
|
| |
It used to be Requires.private, but we expose
harfbuzz types in public headers now.
|
|
|
|
|
|
| |
The correct include is without harfbuzz/.
Closes: https://gitlab.gnome.org/GNOME/pango/issues/387
|
|
|
|
|
| |
Until we move harfbuzz from Requires.private to Requires,
public headers should include harfbuzz/hb.h.
|
|\
| |
| |
| |
| | |
meson: fix darwin_versions
See merge request GNOME/pango!105
|
| |
| |
| |
| | |
The darwin_versions argument should take into account the interface_age
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
We were inserting hyphens after spaces, and in
some other places where they are not desirable.
Fix this by looking at the characters on both
sides of the break when deciding whether to
insert a hyphen.
|