gnome-shell crashed with SIGABRT: mutter:ERROR:../src/core/window.c:4841:set_workspace_state: assertion failed: (workspace == NULL)

[Rocko]

https://errors.ubuntu.com/problem/8317377ffa8b8148baa481d490fd876111eb3c4b

gnome-shell crashes with this bug every time that:

* I am editing a file in geany
* I open the Find window (CTRL-F) and it opens on another monitor
* I type in some un-findable text in the find window and press ENTER (or click the Next button)

The system beeps to indicate that the text isn't found, but then crashes.

Note especially that the find window must be on the other monitor - if it's on the same monitor as geany, gnome-shell doesn't crash.

(Note: I originally thought that if I moved the find window to the other window and then searched, gnome-shell crashed, but I think what happens is a) I open the window, b) I move it, search for something, and it works, c) I close it and re-open it - it then opens on the other monitor and searching will crash gnome-shell.)

gnome-shell restarts, but the crash causes data loss, as most of the running applications don't reappear when it restarts.

In case it's relevant, I have a laptop screen set at 2048x1152 and an external monitor configured to be above it at 2560x1440.

ProblemType: Crash
DistroRelease: Ubuntu 19.10
Package: gnome-shell 3.34.1-1ubuntu1
Uname: Linux 5.4.0-050400rc4-generic x86_64
ApportVersion: 2.20.11-0ubuntu8
Architecture: amd64
CurrentDesktop: ubuntu:GNOME
Date: Tue Oct 22 10:37:36 2019
DisplayManager: gdm3
ExecutablePath: /usr/bin/gnome-shell
ExecutableTimestamp: 1570619894
InstallationDate: Installed on 2019-07-01 (112 days ago)
InstallationMedia: Ubuntu 19.10 "Eoan Ermine" - Alpha amd64 (20190606)
ProcCmdline: /usr/bin/gnome-shell
ProcCwd: /home/rocko
ProcEnviron:
 LANG=en_AU.UTF-8
 LANGUAGE=en_AU:en
 PATH=(custom, no user)
 SHELL=/bin/bash
 XDG_RUNTIME_DIR=<set>
RelatedPackageVersions: mutter-common 3.34.1-1ubuntu1
Signal: 6
SourcePackage: gnome-shell
StacktraceTop:
 () at /lib/x86_64-linux-gnu/libglib-2.0.so.0
 g_assertion_message_expr () at /lib/x86_64-linux-gnu/libglib-2.0.so.0
 () at /lib/x86_64-linux-gnu/libmutter-5.so.0
 () at /lib/x86_64-linux-gnu/libmutter-5.so.0
 () at /lib/x86_64-linux-gnu/libmutter-5.so.0
Title: gnome-shell crashed with SIGABRT
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: sudo
separator:

[Apport retracing service (apport)]

StacktraceTop:
 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50
 __GI_abort () at abort.c:79
 g_assertion_message.cold () from /tmp/apport_sandbox_6jnufst3/usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6200.1
 g_assertion_message_expr () from /tmp/apport_sandbox_6jnufst3/usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6200.1
 set_workspace_state (window=0x558dbf4d7210, on_all_workspaces=1, workspace=0x558dbc9b6b00) at ../src/core/window.c:4841

[Daniel van Vugt (vanvugt)]

Thanks for the bug report. Tracking in: https://errors.ubuntu.com/problem/8317377ffa8b8148baa481d490fd876111eb3c4b

[Rocko (rockorequin)]

Can I provide any other information to help speed this along? This is a pretty severe problem - it's annoying because gnome-shell freezes for a long time before restarting, and then it's extremely annoying because you lose any data in applications that don't survive the restart.

[Daniel van Vugt (vanvugt)]

I can't find any other reports of this crash other than the above link for automatic crash reports. And even then we've only observed the crash 11 times globally, ever. 8 times in eoan and 3 times in bionic.

Sounds like there's something unusual about this machine so please start by uninstalling (don't just disable) these extensions:

'cpufreq@konkor'
'<email address hidden>'
'<email address hidden>'
'<email address hidden>'

[Daniel van Vugt (vanvugt)]

Maybe the usual thing is indeed your workspace/monitor layout.

[Rocko (rockorequin)]

I removed all those extensions (and mate-optimus as well), so ~/.local/share/gnome-shell/extensions/ is completely empty, but the bug is still there.

[Rocko (rockorequin)]

I also removed solaar, which had an indicator (but I think it's showing through app-indicators), with no luck.

I also tried changing the screen configuration to laptop on the left, monitor on the right, and the bug is still present.

[John Agosta (jagosta)]

Regarding comment #10 that this may be this user's specific issue. I have opened bug #1850833 that is classified as a duplicate of this bug. I have also provided in that bug a cleaner path to reproduce.

[John Agosta (jagosta)]

Additional thought on my scenario ... I was also working in a duel monitor mode when gnome crashed. In my case Thunderbird was on my primary monitor, and the thunderbird dialog came up on my secondary monitor. Interacting with this dialog is when gnome crashes.

[Rocko (rockorequin)]

@jagosta: does it only crash if the dialogs are on different monitors? That's what I find.

FWIW, I tried the live USB, but ootb Ubuntu 19.10 doesn't recognise the HDMI monitor, so I can't do dual displays, and I also tried a VirtualBox guest, but even after installing Ubuntu 19.10 and guest drivers, Ubuntu 19.10 doesn't correctly recognise the second monitor, so I can't do dual displays.

[John Agosta (jagosta)]

I have confirmed that it will not crash if I have a single monitor.

In duel monitor mode, i can also reconfirm that after updating (strangely) Chrome browser that the window manager will crash, but will restart -- prior to the update it would crash and not restart.

[Rocko (rockorequin)]

I just discovered that on my laptop this bug only applies to X11 sessions, not to Wayland sessions.

[Rocko (rockorequin)]

I updated the description to note that the text being searched for is not found in the document (which is the case of course for an empty document).

[Jay Cornwall (jcornwall)]

I also see this assertion with a slightly different line number:

[ 4892.053454] desktop gnome-shell[4423]: mutter:ERROR:../src/core/window.c:4849:set_workspace_state: assertion failed: (workspace == NULL)

I installed Pro Motion NG from cosmigo.com through Wine (Ubuntu repository). The crash occurs when opening the "Load Project" file requester.

I have three monitors. The application does not crash if the window is moved to the leftmost or rightmost monitor. It crashes on the center monitor only. This is irrespective of the primary monitor setting, and also of monitor numbering (I changed the physical connections). The resolutions are 1920x1200 (left), 1920x1200 (center), 1920x1080 (right), on an AMD RX 580.

A workaround for this application is to enable Wine's virtual desktop.

[Rocko (rockorequin)]

This is still 100% reproducible with with an up-to-date Ubuntu 20.04.

[Daniel van Vugt (vanvugt)]

Also tracking in: https://errors.ubuntu.com/problem/a3cac4e2a5f6847233c7e4fd6cb99b9428f60ad8

as well the existing: https://errors.ubuntu.com/problem/8317377ffa8b8148baa481d490fd876111eb3c4b

[Rocko (rockorequin)]

Still happens every time in gnome-shell 3.36.0-2ubuntu1 / geany 1.36-1build1.

[Rocko (rockorequin)]

This is fixed for me in gnome-shell 3.36.0-2ubuntu2, thanks!

[Daniel van Vugt (vanvugt)]

Weird. I can't find an exactly matching upstream bug. And the closest fix I can find was mutter!1029 which was released in 3.35.90 already. Regardless, the original reporter says it is fixed AND the ubuntu error tracker says the crash hasn't occurred since version 3.34.1.

If you experience any similar crashes in the near future then those might be bug 1864326 (fixed in 3.36.1 coming soon).

[Rocko (rockorequin)]

Oh, I spoke too soon. I tested it yesterday and it passed, but today it failed. I had geany on the primary monitor this time.

[Daniel van Vugt (vanvugt)]

I would like to verify *which* crash that is. Can you please attach a log:

  journalctl -b0 > journal.txt

and/or identify a crash report link?

Look at https://errors.ubuntu.com/user/ID where ID is the content of file /var/lib/whoopsie/whoopsie-id on the machine. Do you find any links to recent problems on that page? If so then please send the links to us.

[Rocko (rockorequin)]

Do you mean a link to the original crash? This one has the same date as the bug report: https://errors.ubuntu.com/oops/4bb4a564-f475-11e9-b5e5-fa163e102db1

[Daniel van Vugt (vanvugt)]

Thanks. The log in comment #27 doesn't seem to provide clarification other than it *doesn't* look like this bug at all, or any crash for that matter. And the link in comment #28 is not retraceable :(

So I think we were right to call this bug closed in comment #24. But also you seem to be experiencing some other bug so please continue reporting new bugs for new crashes you experience.

[Rocko (rockorequin)]

Do you mean I should recreate this exact same bug and then report it as another one?

[Rocko (rockorequin)]

Ok, I've reported it as bug #1871554.

[Daniel van Vugt (vanvugt)]

Thanks. That's bug 1864326 as I suspected in comment #24.