| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=786887
|
|
|
|
|
|
| |
Skip the space characters in sentence start/end.
https://bugzilla.gnome.org/show_bug.cgi?id=785978
|
|
|
|
|
|
|
| |
Behdad pointed out that I didn't get it right for
flags and for sequences ending in a skin tone modifier.
Fix that.
|
|
|
|
|
|
|
|
| |
We want to delete the entire cluster if it is an Emoji.
To do so, look at prev_GB_type to find Emoji.
This fixes backspacing for sequences like
"man health worker" or "family: man, woman, boy, girl".
|
|
|
|
|
|
|
|
|
|
| |
ZWJ is treated as its own class in TR29, and us lumping
them together and then manually checking wc == 0x200d in
various places was causing us to inadvertedly inserting
grapheme breaks in the middle of Emoji ZWJ sequences
where they are not suppose to be.
Add test cases to verify this.
|
|
|
|
|
| |
Messed up whitespace makes this unwieldy function
even harder to read.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Cairo's CGFont backend already handles scaling fonts to the display
resolution, so scaling to the screen screen resolution in Pango's
CoreText backend generates double-sized text on a Retina display.
The layout test depended on providing a (bogus) resolution of 96DPI and
scaling the font to it in order to obtain the "expected" layout
results. This is replaced by adjusting the font size in test-layout.c
if Pango is compiled for CoreText.
https://bugzilla.gnome.org/show_bug.cgi?id=782393
|
|
|
|
|
|
|
|
|
|
|
|
| |
The FriBiDi code uses #if rather than #ifdef, which pedantic compilers
warn about if the macro is undefined. The new Meson build sets
-Werror=undef, promoting these warnings to failures, and making CI
systems using certain compiler versions (in this case, Debian Jessie)
sad.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
https://bugzilla.gnome.org/show_bug.cgi?id=786192
|
|
|
|
|
| |
This is necessary to make some emoji sequences shape
and render correctly.
|
|
|
|
| |
Patch from Harioum Pratap Singh.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
From a 2014 HarfBuzz commit:
commit 763e5466c0a03a7c27020e1e2598e488612529a7
Author: Behdad Esfahbod <behdad@behdad.org>
Date: Sat Aug 2 16:17:44 2014 -0400
Make it easier to use HB_BUFFER_FLAG_BOT/EOT
Previously, we expected users to provide BOT/EOT flags when the
text *segment* was at paragraph boundaries. This meant that for
clients that provide full paragraph to HarfBuzz (eg. Pango), they
had code like this:
hb_buffer_set_flags (hb_buffer,
(item_offset == 0 ? HB_BUFFER_FLAG_BOT : 0) |
(item_offset + item_length == paragraph_length ?
HB_BUFFER_FLAG_EOT : 0));
hb_buffer_add_utf8 (hb_buffer,
paragraph_text, paragraph_length,
item_offset, item_length);
After this change such clients can simply say:
hb_buffer_set_flags (hb_buffer,
HB_BUFFER_FLAG_BOT | HB_BUFFER_FLAG_EOT);
hb_buffer_add_utf8 (hb_buffer,
paragraph_text, paragraph_length,
item_offset, item_length);
Ie, HarfBuzz itself checks whether the segment is at the beginning/end
of the paragraph. Clients that only pass item-at-a-time to HarfBuzz
continue not setting any flags whatsoever.
Another way to put it is: if there's pre-context text in the buffer,
HarfBuzz ignores the BOT flag. If there's post-context, it ignores
EOT flag.
|
|
|
|
|
|
|
| |
The example code is not correctly escaped, and gtk-doc has become
slightly more strict. This is a good chance to port this bit of the API
reference to Markdown, and avoid all issues in a future without inlined
XML.
|
| |
|
|
|
|
| |
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=785566
|
|
|
|
|
|
| |
Ported from Chromium.
Not hooked yet.
|
|
|
|
|
|
|
| |
Re-write the code for Sentence Boundary,
and use the code style like Grapheme Boundary and Word Boundary.
https://bugzilla.gnome.org/show_bug.cgi?id=782813
|
|
|
|
|
|
|
| |
Improve Word Boundary Rule for Hebrew_Letter, Single_Quote,
Double_Quote and Regional Indicator.
https://bugzilla.gnome.org/show_bug.cgi?id=782813
|
|
|
|
|
|
| |
Support Grapheme Boundaries Rule GB10, GB11, GB12 and GB13.
https://bugzilla.gnome.org/show_bug.cgi?id=782813
|
|
|
|
|
|
|
|
|
|
|
| |
Checked several emoji sequences, the sequence pattern is
like follows:
1. Use zero width joiner to combine two characters with
any width
2. Ignore the width of variation selector, tag and emoji modifier
for the purposes of finding a run or uniform-width characters.
https://bugzilla.gnome.org/show_bug.cgi?id=780669
|
|
|
|
|
|
| |
The keyword is called symbol_prefix, not symbols_prefix.
https://bugzilla.gnome.org/show_bug.cgi?id=785296
|
|
|
|
|
|
|
| |
Arrays can never be NULL, and doing unnecessary checking causes
compilers to show warnings.
https://bugzilla.gnome.org/show_bug.cgi?id=783428
|
|
|
|
|
|
|
|
| |
It seems that it is possible for compilers to use unsigned interger
types to store enum values, so we should cast them to signed interger
types before doing calculation to avoid getting unexpected results.
https://bugzilla.gnome.org/show_bug.cgi?id=783428
|
|
|
|
| |
Owen and I agreed that this doesn't look right.
|
|
|
|
|
|
|
|
|
|
|
| |
When calling into cairo to render glyphs, we want
to use show_text_glyphs and pass the text along,
if the surface supports that operation. But here,
we throw the text away prematurely and end up
always end passing no text or clusters down to
cairo.
https://bugzilla.gnome.org/show_bug.cgi?id=784394
|
|
|
|
| |
s/bytess/bytes/ in a warning.
|
|
|
|
| |
Missing inclusions of freetype via Cairo.
|
|
|
|
|
|
| |
Recently, freetype added checks for various options and forgot to check
all combinations. This breaks building projects including freetype.h
and using -Werror=undef.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
First bug is, we were passing as count to th_brk, the UTF-8
length instead of TIS length. Ouch! I'm not sure how this
was never caught...
The other one was, break-thai was possibly marking a position
is_line_break when break.c has already set it to is_char_break=FALSE.
This broke the invariant that if a position is line-break, then it
must be char-break as well. This, in turn was hitting assertions
in certain conditions. Hit it with this for example:
$ ./pango-view --text 'ส์" (S' --width 43 --font 156px
Note that in a correct world the Latin part of that string should
not reach break-thai.c at all, but that's not how pango-layout.c
breaks right now. See comment before pango_break() call in
pango-layout.c.
|
|
|
|
|
|
|
|
| |
Commit ce097c062ea079bf7888d2b2f1f7b17d34f976b6 was a bit too eager to
mark symbols as public, and marked some internal functions as exported.
If a symbol has an underscore prefix it generally means it's not meant
to be public.
|
|
|
|
|
|
|
|
|
| |
Sub-libraries need to be built after pango-enum-types.h has been
generated; we already have a transitive dependency on the header
through libpango_dep, so we do not need to add the header to the
list of sources.
This avoids symbols leaking in the sub-libraries.
|
|
|
|
|
| |
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=781123
Patch from Takao Fujiwara.
|
|
|
|
|
| |
"soversion:" is for the version placed in the soname. The full library
revision goes into "version:". This affects the symlinks, too.
|
|
|
|
| |
This was an inconsistent mix of tabs and spaces.
|
|
|
|
|
| |
We need to use the `@...@` pattern so that Meson can replace the
symbols.
|
|
|
|
| |
Meson adds it by itself, where appropriate.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Meson is a meta build system that is:
- fast
- simpler to understand and use
- portable to multiple platforms through different backends
- well integrated with the GNOME platform
- well maintained
Using Meson allows us to build Pango much more quickly, and on all the
platforms we currently target, without any loss of functionality,
compared to Autotools.
Some timing comparisons with hot ccache for both build systems:
* autogen.sh: * meson
real 0m11.149s real 0m2.525s
user 0m8.153s user 0m1.609s
sys 0m2.363s sys 0m1.206s
* make -j$(($(nproc) + 2)) * ninja
real 0m9.186s real 0m3.387s
user 0m16.295s user 0m6.887s
sys 0m5.337s sys 0m1.318s
--------------------------------------------------------------
* autotools * meson + ninja
real 0m27.669s real 0m5.772s
user 0m45.622s user 0m8.465s
sys 0m10.698s sys 0m2.357s
System: Intel Core i7-7500U, SSD, 16GB of RAM
|
| |
|
|
|
|
|
|
|
|
| |
It's used in pango-markup.c and pango-color.c, so it should have an
internally-available prototype, instead of an `extern` declaration at
the point of use.
This also avoids a compiler warning for a missing prototype.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
First of all, we need to define `PANGO_UNAVAILABLE`, as it's missing a
definition, and the only reason why nobody caught this is because we're
building without -Wundef, and clearly nobody attempted using
min-required/max-allowed with Pango.
Additionally, we really need a way to disable deprecation warnings from
the compiler.
The complete implementation of the version macros for GLib-based
libraries includes both items above.
|
| |
|
|
|
|
|
| |
The compiler complains that rightmost_space may be used
uninitialized. And it may be right.
|
|
|
|
|
|
| |
Signed-off-by: Philip Withnall <withnall@endlessm.com>
https://bugzilla.gnome.org/show_bug.cgi?id=778663
|
|
|
|
|
|
|
|
|
| |
To make it a bit more explicit that there’s no need to worry about
introspecting them.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
https://bugzilla.gnome.org/show_bug.cgi?id=778663
|
|
|
|
|
|
|
|
|
| |
This should fix a few scanner warnings, but this is not an attempt to
fix them all.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
https://bugzilla.gnome.org/show_bug.cgi?id=778663
|
|
|
|
|
|
|
|
|
|
|
|
| |
The handling for \r or \r\n line endings in pango_read_line() was
broken. It should have discarded the \r or \r\n, but was only doing this
for \n or \n\r. The condition (c == EOF) could never have been reached.
Coverity ID: 1391696
Signed-off-by: Philip Withnall <withnall@endlessm.com>
https://bugzilla.gnome.org/show_bug.cgi?id=778816
|
|
|
|
|
|
|
| |
Improve the behavior of the cairo renderer to take
alpha into account even if no color specified.
https://bugzilla.gnome.org/show_bug.cgi?id=773767
|
|
|
|
|
|
|
| |
Since win32_render_layout and win32_render_layout_line now support
rendering in color, that word no longer applies.
https://bugzilla.gnome.org/show_bug.cgi?id=768679
|
|
|
|
|
|
|
|
| |
Rectangle() uses x1, y1, x2, y2 coordinates instead of x, y, w, h. Fix
the coordinates passed to the function so it draws the background in the
correct location.
https://bugzilla.gnome.org/show_bug.cgi?id=768679
|