| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
| |
Update affected tests
Fixes: #598
|
|
|
|
|
| |
Test that we can round-trip with
pango_layout_x_index (most of the time).
|
|
|
|
|
|
|
| |
Some of these aren't reliable, and were only
added for test coverage.
Fixes: #576
|
|
|
|
|
| |
These cover more of the corner cases fixed in the
previous commits.
|
|
|
|
|
|
|
|
|
|
| |
pango_layout_index_to_line_and_extents needs to
return a run rect even if the line has no non-empty
runs. Failing to do so was causing cursors at the
end of lines to have a height of 0 or 1, which is
not what we want.
Add a testcase.
|
|
|
|
|
| |
Some tests used items from PangoFT2, which may not have been enabled on
platforms where that was optionally built.
|
|
|
|
|
|
| |
Now that we are using run extents for cursor positions,
the vertical extents aren't necessarily constant across
a line anymore. Adapt our assertions.
|
|\
| |
| |
| |
| |
| |
| | |
Reimplement pango_layout_move_cursor_visually
Closes #587, #585, and #157
See merge request GNOME/pango!389
|
| |
| |
| |
| |
| | |
Test the line end handling of
pango_layout_move_cursor_visually.
|
| |
| |
| |
| |
| |
| |
| |
| | |
Reimplement this function based on pango_layout_get_cursor_pos.
This is a bit less efficient, but it fixes cases where the old
implementation would get stuck.
Fixes: #587, #585, #157
|
| |
| |
| |
| | |
Some of these currently fail.
|
| |
| |
| |
| |
| | |
We need to apply trailing properly, by skipping
characters.
|
| |
| |
| |
| |
| |
| |
| | |
Accept values like 200%, in addition to other
ways of specifying sizes in markup.
Fixes: #23
|
| | |
|
|/
|
|
|
|
|
|
| |
Accept values like 12.5pt, in addition to
the other ways of specifying font size in
markup.
Fixes: #67
|
|
|
|
| |
Some of these currently fail.
|
|
|
|
|
|
|
| |
When rendering the text char-by-char, use directional
controls to avoid any bidi reordering.
Update affected tests.
|
|
|
|
|
|
| |
This checks that we get split cursors in some
cases where we expect it. One of the cases was
broken until the previous commit.
|
|
|
|
|
|
|
|
| |
This was quite an odyssee:
- Identify commit from 2008 that caused a regression
- Spot the commit message alluding to a crash
- Identify bug number in NEWS for releases from 2008
- Find the bug in the no longer searchable bugzilla.gnome.org
|
| |
|
|
|
|
|
|
|
|
|
| |
Changes in width only need to break runs in
vertical context.
Update tests and add a new test.
Fixes: #503
|
|
|
|
|
|
|
|
| |
itemize_state_init was setting gravity-related fields
to their initial values too late, after update_attr_iterator
has already set them to the values for the first run.
Test included.
|
|
|
|
| |
This includes stdlib.h for strtol, strtoll and strtod functions declaration
|
|
|
|
|
|
|
|
|
|
| |
pango_attr_iterator_get_attributes only returns
the innermost attribute of each kind. That is correct
for attributes that supersede each other, but not for
font descriptions, which get merged. So, keep all
font attributes in the returned list.
Update affected tests.
|
|
|
|
|
|
| |
For tests where the wrong fonts break things,
check that we have Cantarell, and skip them
if not.
|
|
|
|
|
| |
We were rounding things differently, giving
us off-by-one errors on MacOs.
|
|
|
|
|
| |
The return value of setlocale is only good until
the next call. Also, consistently use en_US.UTF-8.
|
| |
|
| |
|
|
|
|
|
|
|
| |
This is sometimes useful, even if we want to avoid
it in ci. So add an environment variable,
PANGO_TEST_SHOW_FONT to make it show the actual font
instead of 'OMITTED'.
|
|
|
|
|
|
| |
Add more cases to cover move of the branches
in this complicated function. Among other things,
this tests the fix in the previous commit.
|
|
|
|
|
|
|
|
|
|
| |
Make attribute_from_string() parse enum attribute values
in the same way as PangoMarkup, and use it in the
attribute tests.
Still to do: Do the same for flags. The flags parsing
function isn't exported, so this needs our tests to be
statically linked.
|
|
|
|
|
|
|
|
| |
The documentation for this function states that all
bidi control chars are zero width, but it was missing
the Isolate ones. Add them.
Updated tests included.
|
|
|
|
|
| |
This recreates the scenario described in #564.
It works fine.
|
|
|
|
|
|
|
|
| |
Correctly classify tag characters as Grapheme_Extend,
so that we don't end up putting grapheme boundaries
in the middle of tag sequences.
Includes a test.
|
| |
|
|
|
|
|
|
|
| |
Add grapheme boundaries to the output of test-break,
so we can check those easily, too.
Update expected test output for this.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Also add some tests for bidi embedding levels.
|
| |
|
| |
|
|
|
|
|
| |
Avoid g_test_init if we are just generating expected output,
for easier updating.
|
|
|
|
|
|
|
|
| |
Allow setting more layout properties in test-layout,
and check their values after setting them.
Also print out directions and cursor positions.
Update existing tests, and add more.
|
|
|
|
| |
Spot-check a copied layout in test-break.
|
| |
|
| |
|
|
|
|
|
| |
We need space for G_MAXUINT and 4294967285 (in testattributes.c). On
i686, this requires long long, as long is only 32 bits wide.
|