User Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/112.0

Steps to reproduce:

I am using Firefox Snap on Ubuntu with HiDPI. I upgraded Ubuntu from 22.10 to 23.04 (with Gnome 44) and started seeing random crashes, even though Firefox Snap was not updated. Later on, I received a snap update to 112.0.2, but the crashes remained.

The crashes seem to happen without any action from my side, usually at least a few seconds after the last interaction with the UI. I keep seeing 1-2 crash per day. Sorry that I don’t have a more reliable way to reproduce the problem.

I cannot judge whether this is related to #1760001

Actual results:

Firefox silently crashes. No crash reporter window comes up (is this a separate bug?)

I see the following in the systemd journal:
Mai 05 15:04:48 dell gnome-shell[3523]: WL: error in client communication (pid 7292)
Mai 05 15:04:48 dell firefox[7292]: Error 71 (Protocol error) dispatching to Wayland display.
Mai 05 15:04:48 dell firefox_firefox.desktop[7292]: [GFX1-]: Wayland protocol error: wl_surface@77: error 2: Buffer size (1173x128) must be an integer multiple of the buffer_scale (2).
Mai 05 15:04:48 dell firefox_firefox.desktop[7997]: Exiting due to channel error.
Mai 05 15:04:48 dell firefox_firefox.desktop[7756]: Exiting due to channel error.
Mai 05 15:04:48 dell firefox_firefox.desktop[35453]: [GFX1-]: CompositorBridgeChild receives IPC close with reason=AbnormalShutdown
Mai 05 15:04:48 dell systemd[1359]: snap.firefox.firefox.0862649f-ebe5-4909-894f-dab40ddab836.scope: Consumed 1h 9min 24.371s CPU time.

Expected results:

Firefox should not have crashed even on HiDPI.

I think this happens more often while I review pull requests on GitHub, if this helps

Yes, today this happened 4 times in a row while I was reviewing PRs on GitHub

It’s clear from the error message that I use 2x scaling, aka 200% on my desktop. In addition to that, I have the default zoom of 120% in my Firefox.

The Bugbug bot thinks this bug should belong to the 'Core::Widget: Gtk' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Thanks for the report!
This has been fixed in 113 which will be released 2023-05-09.
Until then, you can remove the MOZ_ENABLE_WAYLAND=1 environment variable (which you have manually set).
bug 1813407 landed in 112 and caused bug 1820293, which has been closed as duplicate of bug 1803016, which has been fixed in 113, but not backported to 112.

The crash reporter should have opened.
bug 1826583 comment 40 might have relaxed Wayland crash reporting too intense.

Thank you for the outstanding response, Darkspirit! I’m very impressed and grateful!

:stransky, since you are the author of the regressor, bug 1826583, could you take a look?

For more information, please visit BugBot documentation.

Set release status flags based on info from the regressing bug 1826583

Firefox 113 is now available. Tom, can you confirm that your problem is resolved?

I am having the same issue on Fedora Silverblue 38 running Gnome 44 using the official Flatpak from Flathub.
Here is the console output I got, when a crash happened:

/usr/bin/flatpak run --branch=stable --arch=x86_64 --command=firefox --file-forwarding org.mozilla.firefox

[ERROR glean_core] Error setting metrics feature config: Json(Error("EOF while parsing a value", line: 1, column: 0))
[GFX1-]: Wayland protocol error: wl_surface@132: error 2: Buffer size (1173x128) must be an integer multiple of the buffer_scale (2).

Gdk-Message: 18:33:10.147: Error 71 (Protokolfejl) dispatching to Wayland display.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.
Exiting due to channel error.

Sorry I didn't see that Firefox 113 has been released. Will also test here if it has solved the issue.

The crash should be fixed by bug 1803016 (113). It has been reopened, but it should be closed and a new bug be filed (bug 1803016 comment 26).

I changed this bug's meaning to "Wayland crash occurs but no crash reporter opens". bug 1803016 comment 33 has the same problem.

bug 1653852 will fix Flatpak's crash reporter for crashes that are not caused by Wayland.

In that context I will not be able to confirm whether the crash reporter is fixed or not. Version 113 apparently fixes the only cause for such crashes.

Correct, but please confirm whether the crash has been fixed with 113 or not.

Just for the case can you check Mozilla binaries to make sure it's not snap/flatpak specific issue?
https://fedoraproject.org/wiki/How_to_debug_Firefox_problems#Testing_Mozilla_binaries
Thanks.

I have been using snap version 113 for three days now and have not yet seen a single crash, so the bug looks very much fixed to me. Thanks, everybody!

I guess it does not make sense to test the binaries anymore?

Thanks for the confirmation!

2 bugs have been reported in comment 0:

• crash: bug 1820293 = bug 1803016. Crash fix has been confirmed in comment 16.

• no crash reporter opens: Wayland protocol crashes don't seem to open a crash reporter, Firefox crashes without crash report.
  That's not what was intended: bug 1826583 wanted to put warnings that do not crash into gfx critical error, so that warnings do not crash and crashes crash with crash reporter.

  bug 1832158 (114 Nightly) also has this problem ("I'm unable to get the crash reporter to appear."). This bug has been repurposed for that aspect.

Set release status flags based on info from the regressing bug 1826583

Gian-Carlo, should we prioritize this bug? (lack of a crash reporter, not the crash itself). Thanks.

Please file a new bug, they're cheap. There is no point in trying to repurpose this one and it will just generate confusion.

We should try to check the "Wayland crashes don't generate a crash reporter" outside Flatpak and Snap first, because for example Flatpak crash reporting was known to be broken, so this might already be fixed.

It hasn't been a snap specific issue.

(In reply to Darkspirit from comment #22)

It hasn't been a snap specific issue.

*** This bug has been marked as a duplicate of bug 1803016 ***

it's not specific, but it is good to know for snap-wayland work that there is this issue ...