| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
These were produced with the help of
http://gitlab.gnome.org/matthiasc/layout-editor
|
|
|
|
|
|
| |
Use the new output serialization in test-layout
to store both the input and output in the same
file.
|
|
|
|
|
|
|
| |
If requested, serialize lines, runs, and log attrs.
This will let us use the serialization format to
record not just the test inputs, but outputs as
well.
|
| |
|
|
|
|
|
|
|
|
|
| |
Optionally include context information in the
serialization and restore it. This will be useful
for transporting for giving tests complete data
that affects their output.
Some tests included.
|
|
|
|
|
|
|
| |
This gives us some room to add more things
to these apis, e.g. a way to store not just
the input part of a layout, but also the
output.
|
|
|
|
|
| |
They are produced by deserialization,
so call them that.
|
|
|
|
|
|
|
| |
When looking at scripts, we want to look
at the script of the *previous* character.
And then we need to exclude SHY from the
common script.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When we take a complete item because there are
no breakpoints, we were inserting a hyphen at
the end even though there may not be a breakpoint
there, and ending up with a hyphen in the middle
of an (overlong) line: Brat-wurst.
Test included.
This but was discovered with the help of
http://gitlab.gnome.org/matthiasc/layout-editor
|
|
|
|
|
|
|
|
|
|
|
| |
I went a bit overboard in #603. The vertical bar
is too common, and this usage to specialized to
do it by default. To enable this, we should rather
make log attrs tweakable.
We still replace hyphenation points with hyphens.
Fixes: #630
|
|
|
|
|
| |
Now that we have this api, lets use it for
our tests.
|
|
|
|
|
|
|
|
| |
Add api to serialize PangoLayout, for the benefit
of testing and debugging. Currently, this uses
json, but that is an implementation detail.
Some tests included.
|
| |
|
| |
|
|
|
|
| |
Update affected test outputs.
|
|
|
|
|
|
| |
Add an api to serialize PangoAttrList.
This will be useful in testing and debugging.
|
|
|
|
|
| |
It fails there (although it shouldn't), and I
don't have the means to debug it.
|
| |
|
| |
|
|
|
|
|
| |
Our asan ci is unforgiving, and point out
every memory leak.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Generate random string that looks like a sentence.
* Measure the size at 100 random widths.
* assert that for those widths, when sorted, the measured widths
are increasing and the heights are decreasing.
* Generate random string that looks like a sentence.
* Generate a 2nd string adding hyphen + newline after every char of
original string.
* Assert that wrap-char wrap with width=0 is not larger than the width
of the 2nd string
The tests also generate random strings in LTR, RTL or both at the same
time to spice the test up even more.
It does that by picking random words out of hunspell directories, and if
none are installed, it falls back to lorem ipsum.
|
|
|
|
| |
Some of the layout tests now get better line breaks.
|
| |
|
|
|
|
|
|
|
|
|
| |
We were not passing the correct split_offset to
pango_item_split().
Testcase included.
Fixes: #627
|
| |
|
|
|
|
| |
This checks behavior that is required by GTK.
|
|
|
|
|
|
| |
Test some more tests for roundtripping
casing variants from font description to
font and back.
|
|
|
|
|
|
|
| |
Add a test that verifies that PANGO_VARIANT_SMALL_CAPS
survives a roundtrip from font description to font and
back, and that it shows up in the font as smcp=1 in
the OpenType features.
|
|
|
|
|
| |
I can't debug this using just the ci runner,
so skip it until somebody can.
|
| |
|
|
|
|
| |
This reverts commit 7a4af7a052e7e4187955aad9c4c60853bad806f9.
|
| |
|
|
|
|
|
| |
These need to wait until 20ec670e124e4461075
can be reapplied.
|
| |
|
|\
| |
| |
| |
| |
| |
| | |
Fix empty line heights
Closes #421
See merge request GNOME/pango!481
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When metrics hinting is on, we used to provide rounded
font metrics. We should keep doing that since otherwise
we end up with logical rects sometimes being rounded
and sometimes not, leading to jumping empty lines.
Test included.
Fixes: #421
|
| | |
|
|/ |
|
|
|
|
|
|
|
|
|
|
|
|
| |
We don't really have firm control over the way
attributes are ordered in the list, so the
assumtion that we see proper nesting for baseline
shifts was a bit optimistic.
Just look through all open stack items for a match,
and remove that.
Test included.
|
|\
| |
| |
| |
| | |
Implement baseline shifts
See merge request GNOME/pango!462
|
| |
| |
| |
| | |
A bit more natural than the environment variable.
|
| |
| |
| |
| |
| |
| |
| | |
I added these for improved coverage, but they
get in the way, Should probably just move these
kind of checks to another place and leave this
test focused on dumps.
|
| |
| |
| |
| |
| |
| |
| |
| | |
Use the new baseline-shift and font-scale attributes
to reimplement <sup> and <sub>.
Now they take font metrics into account and nest
properly.
|
| |
| |
| |
| |
| |
| |
| | |
Add a new font-scale attribute to indicate font size
changes due to super- and subscript shifts, and handle
it during item post-processing to find the right font
sizes.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Add a new baseline-shift attribute, which is similar to
rise, but accumulates. In addition, it supports font-
relative values such as superscript and subscript.
We implement support for this by computing baseline
shifts for run during line post-processing, and storing
them in the runs. The renderer now takes these shifts
into account when rendering layout lines.
|
|/ |
|
|
|
|
|
|
| |
We were skipping the test that would have found
the brokenness in the previous fix. Lets see if
we can get the test working on MacOs.
|
| |
|
| |
|