Bug#987947: unblock (pre-approval): gtk+3.0/3.24.24-4

2021-05-04 Thread Sebastian Ramacher 
Control: tags -1 confirmed moreinfo

On 2021-05-02 14:13:56 +0100, Simon McVittie wrote:
> Package: release.debian.org
> Severity: normal
> User: release.debian@packages.debian.org
> Usertags: unblock
> X-Debbugs-Cc: debian-gtk-gn...@lists.debian.org, debian-b...@lists.debian.org
> 
> I'd like to update gtk+3.0 in bullseye to pick up an assortment of fixes
> from upstream. Exactly *which* fixes is something I'm happy to discuss:
> I've prepared a candidate package with a reasonable set. Does this
> look acceptable?

The patches look reasonable to me, so please go ahead and remove the
moreinfo tag once the new version is available in unstable.

Cheers

> 
> unblock gtk+3.0/3.24.24-4
> 
> [ Reason ]
> Backport various targeted fixes from upstream 3.24.25 up to 3.24.29,
> avoiding lower-impact and higher-regression-risk parts (notably avoiding
> theme and input-method changes, which have been more problematic).
> 
> [ Impact ]
> Most of the changes here are about avoiding infrequent crashes, such
> as LP: #1911036 which can happen during drag-and-drop when using the
> proprietary NVIDIA driver. Please let me know if you need more information
> on the impact of any of the crash fixes.
> 
> d/p/x11-Be-quiet-on-exit-by-default.patch silences warnings when a
> GTK application exits in response to the X11 server going away. These
> warnings go to the system log and create a lot of log noise at the
> end of an X11 session (at which time they are completely harmless),
> or if Xorg or Xwayland crashes (in which case we want users to report
> the Xorg/Xwayland crash as a bug, rather than getting distracted by all
> their GTK apps exiting and misinterpreting the log as implying that an
> app crash is the root cause).
> 
> d/p/gdkpixbuf-drawable-Free-the-pixbuf-on-Cairo-error.patch fixes a
> resource leak.
> 
> Fixing LP: #1919404 fixes visual appearance of a specific GNOME Shell
> extension (and anything else with similar behaviour) on HiDPI displays.
> 
> d/p/fontchooser-Fix-some-since-annotations.patch is low-impact but also
> low-risk; it corrects API documentation.
> 
> d/p/placesview-Open-location-even-if-mount-was-not-found.patch fixes
> a regression in Nautilus v40 (not in bullseye), and perhaps other
> file managers. With the gvfs backends for SMB (Samba, CIFS, Windows
> Networking), the intention is that smb://192.168.1.1/ lists the shares
> on 192.168.1.1 and smb://192.168.1.1/music displays a specific share,
> but displaying smb://192.168.1.1/ was not possible. The version of nautilus
> in bullseye is unaffected by this bug.
> 
> [ Tests ]
> Most of the upstream unit tests are run at build-time and as autopkgtests.
> Some tests that are known-broken upstream or are particularly sensitive
> to implementation details of other packages are skipped.
> 
> Manual testing: I use GNOME daily, and have been using the newer upstream
> releases in experimental, from which these changes were backported. I've
> now switched to this proposed version to get it some more testing. Some
> of the fixes here are also used in Ubuntu.
> 
> I have a machine with the NVIDIA driver which exhibits the warnings
> silenced by d/p/x11-Be-quiet-on-exit-by-default.patch during session
> logout, and I confirm that the patch does successfully silence them.
> 
> I have not attempted to reproduce the various crashes.
> 
> [ Risks ]
> This is obviously an important key package that lots of things depend on.
> Technically it also has a udeb, although I'm fairly sure d-i is still
> using GTK 2 and so the udeb is not actually used for anything yet.
> 
> The changes are targeted and simple (avoiding NULL dereferences and that
> sort of thing).
> 
> I'm not particularly attached to any of the specific fixes here, and
> I'm happy to drop them or postpone them to bullseye-pu if the release
> team consider them to be an unacceptable risk. I don't think any of the
> changes have dependencies between them, so we can keep or drop any
> particular change as needed.
> 
> [ Checklist ]
>   [x] all changes are documented in the d/changelog
>   [x] I reviewed all changes and I approve them
>   [x] attach debdiff against the package in testing

> diffstat for gtk+3.0-3.24.24 gtk+3.0-3.24.24
> 
>  changelog   |   
> 43 +
>  patches/Fix-a-possible-crash-in-gtk_show_uri.patch  |   
> 27 +
>  patches/Wayland-ignore-touch-tablet-events-on-destroyed-surfaces.patch  |  
> 241 ++
>  patches/cssshadowvalue-Apply-device-scale-to-the-offset-when-blur.patch |   
> 30 +
>  patches/fontchooser-Fix-some-since-annotations.patch|   
> 44 +
>  patches/gdkpixbuf-drawable-Free-the-pixbuf-on-Cairo-error.patch |   
> 23 
>  patches/label-Skip-updating-link-state-if-we-have-no-layout.patch   |   
> 31 +
>  patches/placesview-Open-location-even-if-mount-was-not-found.patch  |   
> 46 +
>  patches/scale-Fix-sporadic-criticals.patch

Bug#987947: unblock (pre-approval): gtk+3.0/3.24.24-4

2021-05-02 Thread Cyril Brulebois 
Hi Simon,

Simon McVittie  (2021-05-02):
> [ Risks ]
> This is obviously an important key package that lots of things depend
> on. Technically it also has a udeb, although I'm fairly sure d-i is
> still using GTK 2 and so the udeb is not actually used for anything
> yet.

That's absolutely correct: no objections to anything you would like to
update on your side.


Cheers,
-- 
Cyril Brulebois (k...@debian.org)
D-I release manager -- Release team member -- Freelance Consultant


signature.asc
Description: PGP signature

Bug#987947: unblock (pre-approval): gtk+3.0/3.24.24-4

2021-05-02 Thread Simon McVittie 
Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock
X-Debbugs-Cc: debian-gtk-gn...@lists.debian.org, debian-b...@lists.debian.org

I'd like to update gtk+3.0 in bullseye to pick up an assortment of fixes
from upstream. Exactly *which* fixes is something I'm happy to discuss:
I've prepared a candidate package with a reasonable set. Does this
look acceptable?

unblock gtk+3.0/3.24.24-4

[ Reason ]
Backport various targeted fixes from upstream 3.24.25 up to 3.24.29,
avoiding lower-impact and higher-regression-risk parts (notably avoiding
theme and input-method changes, which have been more problematic).

[ Impact ]
Most of the changes here are about avoiding infrequent crashes, such
as LP: #1911036 which can happen during drag-and-drop when using the
proprietary NVIDIA driver. Please let me know if you need more information
on the impact of any of the crash fixes.

d/p/x11-Be-quiet-on-exit-by-default.patch silences warnings when a
GTK application exits in response to the X11 server going away. These
warnings go to the system log and create a lot of log noise at the
end of an X11 session (at which time they are completely harmless),
or if Xorg or Xwayland crashes (in which case we want users to report
the Xorg/Xwayland crash as a bug, rather than getting distracted by all
their GTK apps exiting and misinterpreting the log as implying that an
app crash is the root cause).

d/p/gdkpixbuf-drawable-Free-the-pixbuf-on-Cairo-error.patch fixes a
resource leak.

Fixing LP: #1919404 fixes visual appearance of a specific GNOME Shell
extension (and anything else with similar behaviour) on HiDPI displays.

d/p/fontchooser-Fix-some-since-annotations.patch is low-impact but also
low-risk; it corrects API documentation.

d/p/placesview-Open-location-even-if-mount-was-not-found.patch fixes
a regression in Nautilus v40 (not in bullseye), and perhaps other
file managers. With the gvfs backends for SMB (Samba, CIFS, Windows
Networking), the intention is that smb://192.168.1.1/ lists the shares
on 192.168.1.1 and smb://192.168.1.1/music displays a specific share,
but displaying smb://192.168.1.1/ was not possible. The version of nautilus
in bullseye is unaffected by this bug.

[ Tests ]
Most of the upstream unit tests are run at build-time and as autopkgtests.
Some tests that are known-broken upstream or are particularly sensitive
to implementation details of other packages are skipped.

Manual testing: I use GNOME daily, and have been using the newer upstream
releases in experimental, from which these changes were backported. I've
now switched to this proposed version to get it some more testing. Some
of the fixes here are also used in Ubuntu.

I have a machine with the NVIDIA driver which exhibits the warnings
silenced by d/p/x11-Be-quiet-on-exit-by-default.patch during session
logout, and I confirm that the patch does successfully silence them.

I have not attempted to reproduce the various crashes.

[ Risks ]
This is obviously an important key package that lots of things depend on.
Technically it also has a udeb, although I'm fairly sure d-i is still
using GTK 2 and so the udeb is not actually used for anything yet.

The changes are targeted and simple (avoiding NULL dereferences and that
sort of thing).

I'm not particularly attached to any of the specific fixes here, and
I'm happy to drop them or postpone them to bullseye-pu if the release
team consider them to be an unacceptable risk. I don't think any of the
changes have dependencies between them, so we can keep or drop any
particular change as needed.

[ Checklist ]
  [x] all changes are documented in the d/changelog
  [x] I reviewed all changes and I approve them
  [x] attach debdiff against the package in testing
diffstat for gtk+3.0-3.24.24 gtk+3.0-3.24.24

 changelog   |   43 +
 patches/Fix-a-possible-crash-in-gtk_show_uri.patch  |   27 +
 patches/Wayland-ignore-touch-tablet-events-on-destroyed-surfaces.patch  |  241 ++
 patches/cssshadowvalue-Apply-device-scale-to-the-offset-when-blur.patch |   30 +
 patches/fontchooser-Fix-some-since-annotations.patch|   44 +
 patches/gdkpixbuf-drawable-Free-the-pixbuf-on-Cairo-error.patch |   23 
 patches/label-Skip-updating-link-state-if-we-have-no-layout.patch   |   31 +
 patches/placesview-Open-location-even-if-mount-was-not-found.patch  |   46 +
 patches/scale-Fix-sporadic-criticals.patch  |   52 ++
 patches/series  |   11 
 patches/updateiconcache-Sort-list-of-entries.patch  |1 
 patches/x11-Be-quiet-on-exit-by-default.patch   |   49 ++
 patches/x11-Don-t-beep-on-untrusted-displays.patch  |   42 +
 patches/x11-dnd-Ignore-XErrors-from-the-COW.patch   |   36 +
 14 files changed, 676