~vcs-imports/gtk/+git/gtk:fix-subpixel-positions

Last commit made on 2024-02-17
Get this branch:
git clone -b fix-subpixel-positions https://git.launchpad.net/~vcs-imports/gtk/+git/gtk

Branch merges

Branch information

Name:
fix-subpixel-positions
Repository:
lp:~vcs-imports/gtk/+git/gtk

Recent commits

27c978c... by Matthias Clasen <email address hidden>

gsk: Fix ink leakage in node diffing

We do need to enlarge the node bounds for text nodes after all,
to prevent leaking ink, in particular with fractional scaling.

This reverts 24de5ffd4eac7f3de4234.

add0afa... by Matthias Clasen <email address hidden>

gsk: Fix subpixel positioning

The previous code did not subtract the fractional part from the
origin after determining it, and it neglected the case where we
are close enough to the next integral position to just use that.

ad719a9... by Matthias Clasen <email address hidden>

Merge branch 'ink-rect-fixaroo' into 'main'

gsk: Stop padding text node bounds

See merge request GNOME/gtk!6912

0c4ea9b... by Benjamin Otte <email address hidden>

Merge branch 'wip/otte/for-main' into 'main'

gpu: External textures are never mipmap'able

See merge request GNOME/gtk!6911

24de5ff... by Matthias Clasen <email address hidden>

gsk: Stop padding text node bounds

This should not be necessary, provided that the ink extents that
pango provides are accurate.

Update affected tests.

4933bc5... by Benjamin Otte <email address hidden>

gpu: External textures are never mipmap'able

We were just assuming they were if the format matches.

Fixes crashes in Webkit where the external texture is actually a dmabuf
imported as an EGL image.

ecbee35... by Matthias Clasen <email address hidden>

Merge branch 'matthiasc/for-main' into 'main'

gtk-demo: Add keywords to some demos

See merge request GNOME/gtk!6910

5552b5c... by Matthias Clasen <email address hidden>

ci: Add another xfail

This is reported in
https://gitlab.gnome.org/GNOME/gtk/-/issues/6450

2baa899... by Matthias Clasen <email address hidden>

NEWS: Add a new about mesa versions

aaaddd1... by Matthias Clasen <email address hidden>

testsuite: Mark the randomized path tests as flaky

They are, and having them run in ci is counterproductive for now.