gtk4_4.14.4+ds-5_source.changes ACCEPTED into unstable

Wed Jul 31 19:50:09 BST 2024

Thank you for your contribution to Debian.

Accepted:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Format: 1.8
Date: Wed, 31 Jul 2024 14:30:30 -0400
Source: gtk4
Built-For-Profiles: noudeb
Architecture: source
Version: 4.14.4+ds-5
Distribution: unstable
Urgency: medium
Maintainer: Debian GNOME Maintainers <pkg-gnome-maintainers at lists.alioth.debian.org>
Changed-By: Jeremy Bícha <jbicha at ubuntu.com>
Closes: 1072395 1077181 1077192 1077287 1077288 1077289
Changes:
 gtk4 (4.14.4+ds-5) unstable; urgency=medium
 .
   [ Simon McVittie ]
   * d/rules: Add bug reference for #1077289
   * d/changelog: Previous release only mitigated #1077178 rather than
     fixing it
 .
   [ Jeremy Bícha ]
   * Release to unstable (Closes: #1072395)
 .
 gtk4 (4.14.4+ds-4) experimental; urgency=medium
 .
   [ Matthias Geiger ]
   * d/rules: Skip label-shadows reftest on mips64el and riscv64.
     The rendering differences are minor enough to not reflect a real problem
     for applications. (Mitigates: #1077178)
 .
   [ Simon McVittie ]
   * d/patches: Improve patch metadata
   * d/patches: Rearrange patch series, closest-to-upstream first.
     No functional change.
   * d/p/tests-Fix-signature-of-GtkTestATContext-update-selection-.patch:
     Add patch to fix a11y/text, a11y/textview tests instead of skipping them
     (Closes: #1077181)
   * d/p/gskpathop-Introduce-a-type-to-represent-an-aligned-graphe.patch:
     Align GskPath points to an 8-byte boundary where necessary
     (Closes: #1077192, #1077288)
   * d/rules: Skip headless test suite explicitly, rather than via a patch.
     It's equally easy to skip these tests this way, and less delta vs.
     upstream in our patch series makes it easier to manage.
   * d/rules: Reduce test suites specifically skipped on big-endian.
     We now skip these test suites unconditionally on all architectures,
     because upstream say they are very sensitive to small behaviour
     changes in the GL driver. Given that, there's no need to skip them a
     second time on big-endian systems.
   * d/p/gskvulkanmemory.c-Use-VK_NULL_HANDLE-for-VkDeviceMemory.patch,
     d/p/gdkvulkancontext.c-Use-pointers-to-hash-VkShaderModule.patch,
     d/p/gskvulkandevice.c-Put-Vk-Pipeline-RenderPass-in-structure.patch:
     Add patches from upstream 4.14.x branch to fix build on i386.
     The Vulkan renderer in the previous i386 package built successfully with
     gcc 13, but probably didn't actually work. gcc 14 is stricter about
     implicit integer/pointer conversions, so it failed to build.
   * d/rules: Build with -ffloat-store on i386.
     Some of the tests (and more in 4.15.x) fail if this is not done,
     because i387 excess precision results in numeric values being
     calculated differently. Because few GTK developers run it on i386
     (and in particular, upstream do not), running the test suite and not
     skipping more of it than strictly necessary is important, because
     that's the only mechanism available to discover whether the library
     has regressed on less-used architectures.
     (Closes: #1077289)
   * d/p/gdk-Check-some-assertions-on-our-architectures.patch:
     Check some assertions for the GTK upstream maintainers.
     This has no runtime impact, but will fail the build if some reasonable
     assumptions turn out to be untrue.
   * d/control: Depend on libgles2.
     This is used by default since 4.14. It isn't strictly mandatory because
     another backend can be chosen via environment variables, but the
     failure mode if it is missing is very bad (applications crash) so it
     seems proportionate to make it a Depends. (Closes: #1077287)
 .
 gtk4 (4.14.4+ds-3) experimental; urgency=medium
 .
   * Extra workarounds for tests issues in the new version
   * debian/ignore.keyfile:
     - increase tolerated-diff-pixels to workaround some reftests still
       failing the build on s390x despite being in the ignore list
   * debian/patches/ignore_a11ytext_i386.patch:
     - ignore a11y/text and a11y/textview tests on i386 for now, they
       are failing on the builders for some reason
   * d/p/d/tests-Mark-memorytexture-as-expected-to-fail-on-big-endian.patch:
     - mark also memorytexture as expected to fail on big endian
   * debian/rules:
     - disable gsk tests on i386 and arm since they are currently buggy
     - ignore extra reftests that are now failing on s390x
 .
 gtk4 (4.14.4+ds-2) experimental; urgency=medium
 .
   * Increase general build test timeout. Thanks to Matthias Geiger.
 .
 gtk4 (4.14.4+ds-1) experimental; urgency=medium
 .
   * New upstream release
 .
 gtk4 (4.14.3+ds-1) experimental; urgency=medium
 .
   * New upstream release
   * Switch from cups backend to cpdb
 .
 gtk4 (4.14.2+ds-1) experimental; urgency=medium
 .
   * New upstream release. Upload to experimental but the tests situation
     is not ideal. Skip some of testsuites to allow building since it seems
     the gsk tests can hit rounding differences depending of the renderer
     in use an the software rendering stack version and build options.
     We saw tests issue also on Ubuntu for s390x and 32 bits architectures
     but I'm not applying workaround for those in Debian yet because they
     are somehow wrong (especially on 32 bits which seems to have real
     upstream issues) and to see if we get the same errors on the Debian
     builders.
   * Update the packaging to no refer to the ffmpeg experimental backend
     which has been removed upstream now
   * debian/rules:
     - skip the gsk compare suites those seem to be giving inconsistant
       results depending on the rendering stack
     - update for the new convert-emoji syntax
 .
   [ Nathan Pratta Teodosio ]
   * debian/rules:
     - Make Bash scripts in installed-tests/gtk-4.0/tools/ executable
       to fix autopkgtests.
 .
   [ Jeremy Bícha ]
   * New upstream release
Checksums-Sha1:
 6bf191de2bf636648887a75cbbbb8dcf386fb08c 4779 gtk4_4.14.4+ds-5.dsc
 f56c14eb5abe4518d44194579e2db27905be3792 3612628 gtk4_4.14.4+ds-5.debian.tar.xz
 f7909e79bbf2bcfe0fb4c690f378db670cf86ba7 17466 gtk4_4.14.4+ds-5_source.buildinfo
Checksums-Sha256:
 0b95f8bcd90be3ba6d8eec27f8d324f34b3ba78d5c195c77ac7406e366ff738f 4779 gtk4_4.14.4+ds-5.dsc
 27ca4a16604c772520b34206772d46c7d115d2ac8af4e621c3f396bb89410e3a 3612628 gtk4_4.14.4+ds-5.debian.tar.xz
 4b493ac265032db51c90e7ff593ca4f44f4fb9a7db5595bc78e457fc36d7baa7 17466 gtk4_4.14.4+ds-5_source.buildinfo
Files:
 f95621ea1ff89803e11a6652649cd3d9 4779 libs optional gtk4_4.14.4+ds-5.dsc
 ead9a28820c651b6a7b09c1f88f9face 3612628 libs optional gtk4_4.14.4+ds-5.debian.tar.xz
 600ec6da4d770a252470c1c4d8299f64 17466 libs optional gtk4_4.14.4+ds-5_source.buildinfo

-----BEGIN PGP SIGNATURE-----

iQIzBAEBCgAdFiEETQvhLw5HdtiqzpaW5mx3Wuv+bH0FAmaqhEkACgkQ5mx3Wuv+
bH1C1xAAjBWLJJjjKtH7wN28PEaRXfuL9prZ33jAY10jeeuMQ1icf+RM2YmnRbyG
Wd8AujC8yaPmd5/+KUdWYCJLuOIx/Z93vrLzc8BMAsOwXXa6QntaRD/gkoSHkkYh
am3Kuw4D/lDZpMhABJ9VocV/D5FWhZQ4D1vvkQQKMLvkVG5A6aPjl4kH7YMQk4y9
HfeCgO5zmk2FYvoBS/XIe8v6p9hTV7Wx8YCA50Se7svnubNspWe5gMnrtxLiOU7u
miJJMCK5clcL6A3e/k37/fbRWXbc62hklsfXnnrQMGqI0UZu4go5t4LRp8jp1yD0
Oj7oAwiXqDAvmnCiH1/qxbIm2dCgQ1h7YAcA38sRpzSZdKiFfeqS5x3+D2t5h/7G
gpCC7FJQ31itmsmtFXbBoLhgP4l5vlAnuxJruBRSUJ5M4GFS8tHuwcXl2pJyZTCm
rL78J93ByArISiRuoQmw5J0iB/wBpHSmYtcvU6gI/LEIXNRWkzTZcuvRGQkjZegJ
b219fQmq4Z5DWHhrM6sO8r8mMeORfRufmcxHr6zw1XDqBHWXvFMQ3Ti4luaPaynF
FRDglduf1n6dsKQvgJ90lwLMhpDMXxOV2RT1/7hy+KNFrkvBqrMijcmfzqUcqP0u
T73YHW3VuyoELrL9OIGMo5JAiVS87DseI4ghR3Vrv26JDYTFClo=
=E3Rh
-----END PGP SIGNATURE-----

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
URL: <http://alioth-lists.debian.net/pipermail/pkg-gnome-maintainers/attachments/20240731/472fbb69/attachment.sig>