Unfortunately, it seems that the fix for Bug #1819195 was not enough. I still have occasional crashes, although not that often. One happened today: https://crash-stats.mozilla.org/report/index/edb65646-bc3b-484f-90a3-4cf820200429 It was (as always) during BlueJeans meeting, when I was talking :/ ~~~ $ rpm -q firefox firefox-75.0-2.fc33.x86_64 ~~~
It looks like https://bugzilla.mozilla.org/show_bug.cgi?id=1628583 which may be a bit difficult to backport :(
Can you try to use Firefox nightly with Wayland for the BJ session and check if you see any crash there? Thanks.
Yet another appearance of the issue: https://crash-stats.mozilla.org/report/index/aade3128-d516-49b8-b1dc-3f8070200512 ~~~ $ rpm -q firefox firefox-76.0-2.fc33.x86_64 ~~~
Yes, Bug 1628583 is fixed in FF77.
Firefox 77 is coming next week but can you try to enable webrender? It should solve the https://crash-stats.mozilla.org/report/index/aade3128-d516-49b8-b1dc-3f8070200512 Go to about:config and set gfx.webrender.enabled to true.
(In reply to Martin Stransky from comment #5) > Go to about:config and set gfx.webrender.enabled to true. Does this need FF restart? Will this option become default in FF 77?
(In reply to Vít Ondruch from comment #6) > (In reply to Martin Stransky from comment #5) > > Go to about:config and set gfx.webrender.enabled to true. > > Does this need FF restart? Yes. > Will this option become default in FF 77? No, it enrolls slowly, see https://wiki.mozilla.org/Platform/GFX/WebRender_Where
I don't thin I have met the issue in recent days (using FF 77 since June 9). Instead, I hit today another issue (bug 1847005).
*** Bug 1847005 has been marked as a duplicate of this bug. ***
It's the same bug but with a different bt. 78 should fix that, it comes out 30.6.
Is this something different? https://crash-stats.mozilla.org/report/index/5bff3195-2fd8-4245-ba3f-826990200713 This happened with firefox-78.0-1.fc33.x86_64
And also with firefox-78.0.2-1.fc33.x86_64: https://crash-stats.mozilla.org/report/index/69219597-e8dd-4c8c-9541-feb5d0200715
The previous crashes typically happened during confcalls via Bluejeans. But the following happened when I was typing into the search box of "Simple Tab Groups" extension: https://crash-stats.mozilla.org/report/index/0c5bbfb0-b121-498f-bea3-b43fa0200715
(In reply to Vít Ondruch from comment #13) > The previous crashes typically happened during confcalls via Bluejeans. But > the following happened when I was typing into the search box of "Simple Tab > Groups" extension: > > https://crash-stats.mozilla.org/report/index/0c5bbfb0-b121-498f-bea3- > b43fa0200715 This is fixed by https://bugzilla.mozilla.org/show_bug.cgi?id=1641778 which comes in Firefox 79 next week.
This bug appears to have been reported against 'rawhide' during the Fedora 33 development cycle. Changing version to 33.
It seems that the situation with FF 79 is better then it used to be, but: https://crash-stats.mozilla.org/report/index/f58e44ca-de4c-4875-9c04-e0f300200904
Will look at it, Thanks.
btw. Can you please attach content of about:support? Thanks.
Created attachment 1714727 [details] about:support content
Do you still see the crashes with Firefox 81? Thanks.
I am still on firefox-79.0-5.fc32.x86_64 (because it seems you had issue building FF for Rawhide for some while), but looking at about:crashes it seems the last crash I observed was the one I referred above on Sept 4. Going to update to FF81 out of curiosity.
Thanks for testing. 81 packages for Fedora 33 are available here: https://koji.fedoraproject.org/koji/buildinfo?buildID=1614366
I have updated to: ~~~ $ rpm -q firefox firefox-81.0-6.fc34.x86_64 ~~~ and I'll be back in case of troubles.
New crash today: https://crash-stats.mozilla.org/report/index/7cc3fb42-7072-49a9-833f-3a0f00201007
(In reply to Vít Ondruch from comment #24) > New crash today: > > https://crash-stats.mozilla.org/report/index/7cc3fb42-7072-49a9-833f- > 3a0f00201007 Can you please try the latest packages - firefox-81.0.1-5.fc33 - from https://koji.fedoraproject.org/koji/buildinfo?buildID=1621239 ? It should contain a fix for this one, https://bugzilla.mozilla.org/show_bug.cgi?id=1656727 Thanks.
Updated. I'll keep watching. ~~~ $ rpm -q firefox firefox-81.0.1-5.fc34.x86_64 ~~~
Unfortunately a patch from https://bugzilla.mozilla.org/show_bug.cgi?id=1656727 causes a regression https://bugzilla.redhat.com/show_bug.cgi?id=1886243 so I'm going to revert it in next update firefox-81.0.1-7. I'll look at the patch and try to fix that.
(In reply to Martin Stransky from comment #27) I have already commented there, but if I should choose, I prefer the funky tab popup (which I don't know if I even use) to FF crashes. But of course, the bug 1886243 refers to F32 while I am on Rawhide, which might make a difference.
New builds with a fix for this one are on the way - https://koji.fedoraproject.org/koji/buildinfo?buildID=1621842
err, rawhide builds are here - https://koji.fedoraproject.org/koji/buildinfo?buildID=1621843
Today I had strange failure: ~~~ $ journalctl -b | grep firefox ... snip ... Oct 14 16:00:38 localhost.localdomain firefox.desktop[1648975]: [GFX1-]: Receive IPC close with reason=AbnormalShutdown Oct 14 16:00:38 localhost.localdomain firefox.desktop[1653738]: Exiting due to channel error. Oct 14 16:00:38 localhost.localdomain firefox.desktop[1648975]: Exiting due to channel error. Oct 14 16:00:38 localhost.localdomain firefox.desktop[1701097]: [GFX1-]: Receive IPC close with reason=AbnormalShutdown Oct 14 16:00:38 localhost.localdomain firefox.desktop[1701052]: [GFX1-]: Receive IPC close with reason=AbnormalShutdown Oct 14 16:00:38 localhost.localdomain firefox.desktop[1644803]: Exiting due to channel error. Oct 14 16:00:38 localhost.localdomain firefox.desktop[1700892]: [GFX1-]: Receive IPC close with reason=AbnormalShutdown Oct 14 16:00:38 localhost.localdomain firefox.desktop[1701052]: Exiting due to channel error. Oct 14 16:00:38 localhost.localdomain firefox.desktop[1700892]: Exiting due to channel error. Oct 14 16:00:38 localhost.localdomain firefox.desktop[1701097]: Exiting due to channel error. Oct 14 16:00:38 localhost.localdomain firefox.desktop[1641125]: Exiting due to channel error. Oct 14 16:00:38 localhost.localdomain firefox.desktop[1641083]: Exiting due to channel error. Oct 14 16:00:38 localhost.localdomain firefox.desktop[1644157]: Exiting due to channel error. Oct 14 16:00:38 localhost.localdomain firefox.desktop[1642803]: Exiting due to channel error. Oct 14 16:00:42 localhost.localdomain firefox.desktop[2021274]: restorecon: Could not set context for /home/vondruch/.mozilla/firefox/13qjb9o9.default/cookies.sqlite-wal: No such file or directory Oct 14 16:00:42 localhost.localdomain rtkit-daemon[823]: Successfully made thread 2021362 of process 2021268 (/usr/lib64/firefox/firefox) owned by '16025' RT at priority 10. Oct 14 16:00:43 localhost.localdomain systemd[1694]: app-gnome-firefox-1640930.scope: Succeeded. Oct 14 16:00:43 localhost.localdomain systemd[1694]: app-gnome-firefox-1640930.scope: Consumed 1d 1h 43min 12.127s CPU time. ... snip ... $ rpm -q firefox firefox-81.0.1-5.fc34.x86_64 ~~~ I might just suspect, that is again the same error, because there is no report in `about:crashes` neither ABRT caught anything :/
Actually this was preceded by this: ~~~ Oct 14 16:00:38 localhost.localdomain gnome-shell[1832]: WL: error in client communication (pid 1640930) Oct 14 16:00:38 localhost.localdomain MainThread[1640930]: Error flushing display: Roura přerušena (SIGPIPE) ~~~
Do you have anything at about:crashes? ABRT ignores Firefox failures as we use upstream tracker. Thanks.
As I said, there is nothing in about:crashes. It just went away without any notice except journal.
(In reply to Vít Ondruch from comment #34) > As I said, there is nothing in about:crashes. It just went away without any > notice except journal. I see. In that case it looks like Firefox quit itself after the error, that perhaps come from gtk.
Not sure this is the same, but the crash is again in wayland-client: https://crash-stats.mozilla.org/report/index/314a6d21-41d9-4178-bdad-18e830201021 ~~~ $ rpm -q firefox firefox-82.0-5.fc33.x86_64 ~~~
We need to wait until debug symbols are loaded at mozilla servers, that's done when we ship the builds as Fedora updates.
(In reply to Vít Ondruch from comment #32) > Actually this was preceded by this: > > ~~~ > Oct 14 16:00:38 localhost.localdomain gnome-shell[1832]: WL: error in client > communication (pid 1640930) > Oct 14 16:00:38 localhost.localdomain MainThread[1640930]: Error flushing > display: Roura přerušena (SIGPIPE) > ~~~ Looks like a duplicate of Bug 1888920. *** This bug has been marked as a duplicate of bug 1888920 ***