Is it possible to add fatal issue 1633365 to the list of blocker bugs for F30?
Almost everyone running F28 and F29 Gnome desktop worldwide has reported some variant of this very high severity evolution bug that easily causes a full desktop crash. I believe that its time to get this bug to block the next release.
On Thu, Mar 21, 2019 at 6:34 PM, John Mellor john.mellor@gmail.com wrote:
Is it possible to add fatal issue 1633365 to the list of blocker bugs for F30?
Almost everyone running F28 and F29 Gnome desktop worldwide has reported some variant of this very high severity evolution bug that easily causes a full desktop crash. I believe that its time to get this bug to block the next release.
Evolution bugs cannot be considered blockers, because Evolution is no longer installed by default.
But you linked to a gnome-shell crash, not an Evolution bug. gnome-shell crashes could absolutely be blockers.
But the crash is in X11-specific code, so it can't affect our default user experience, so that's a pretty strong argument against making it a blocker. It's also not a regression from F29. Did you try using Wayland?
Michael
On Thu, 2019-03-21 at 18:58 -0500, mcatanzaro@gnome.org wrote:
On Thu, Mar 21, 2019 at 6:34 PM, John Mellor john.mellor@gmail.com wrote:
Is it possible to add fatal issue 1633365 to the list of blocker bugs for F30?
Almost everyone running F28 and F29 Gnome desktop worldwide has reported some variant of this very high severity evolution bug that easily causes a full desktop crash. I believe that its time to get this bug to block the next release.
Evolution bugs cannot be considered blockers, because Evolution is no longer installed by default.
But you linked to a gnome-shell crash, not an Evolution bug. gnome-shell crashes could absolutely be blockers.
But the crash is in X11-specific code, so it can't affect our default user experience, so that's a pretty strong argument against making it a blocker. It's also not a regression from F29. Did you try using Wayland?
My F30 login options are "Gnome", "Gnome Classic" and "Gnome on Xorg". I am always using "Gnome", so I believe that I am normally on Wayland, not X11. The bug has existed in F29 and F30, but noes not happen in earlier releases.
Hello, please report the bug in the Bugzilla and provide any info you have so that we could reproduce it. You can definitely propose it as a blocker in the blocker bugs application ( https://qa.fedoraproject.org/blockerbugs/propose_bug ). Proposals are discussed every Monday on a Blocker Bug meeting (usually at 17 UTC), where you can come and have your say in the vote.
Lukas
On Fri, Mar 22, 2019 at 2:25 AM John Mellor john.mellor@gmail.com wrote:
On Thu, 2019-03-21 at 18:58 -0500, mcatanzaro@gnome.org wrote:
On Thu, Mar 21, 2019 at 6:34 PM, John Mellor john.mellor@gmail.com wrote:
Is it possible to add fatal issue 1633365 to the list of blocker bugs for F30?
Almost everyone running F28 and F29 Gnome desktop worldwide has reported some variant of this very high severity evolution bug that easily causes a full desktop crash. I believe that its time to get this bug to block the next release.
Evolution bugs cannot be considered blockers, because Evolution is no longer installed by default.
But you linked to a gnome-shell crash, not an Evolution bug. gnome-shell crashes could absolutely be blockers.
But the crash is in X11-specific code, so it can't affect our default user experience, so that's a pretty strong argument against making it a blocker. It's also not a regression from F29. Did you try using Wayland?
My F30 login options are "Gnome", "Gnome Classic" and "Gnome on Xorg". I am always using "Gnome", so I believe that I am normally on Wayland, not X11. The bug has existed in F29 and F30, but noes not happen in earlier releases.
desktop mailing list -- desktop@lists.fedoraproject.org To unsubscribe send an email to desktop-leave@lists.fedoraproject.org Fedora Code of Conduct: https://getfedora.org/code-of-conduct.html List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/desktop@lists.fedoraproject.or...
On Thu, Mar 21, 2019 at 8:25 PM, John Mellor john.mellor@gmail.com wrote:
My F30 login options are "Gnome", "Gnome Classic" and "Gnome on Xorg". I am always using "Gnome", so I believe that I am normally on Wayland, not X11. The bug has existed in F29 and F30, but noes not happen in earlier releases.
Something is wrong then, because in that crash gnome-shell is running as an X11 client. It's receiving an X11 protocol error from the X server. These are fun because they're asynchronous, so the backtrace doesn't actually include enough diagnostic information to do anything about it. But if it really happens 50% of the time, then hopefully a developer would be able to reproduce....
Michael
On Fri, 2019-03-22 at 08:41 -0500, mcatanzaro@gnome.org wrote:
On Thu, Mar 21, 2019 at 8:25 PM, John Mellor john.mellor@gmail.com wrote:
My F30 login options are "Gnome", "Gnome Classic" and "Gnome on Xorg". I am always using "Gnome", so I believe that I am normally on Wayland, not X11. The bug has existed in F29 and F30, but noes not happen in earlier releases.
Something is wrong then, because in that crash gnome-shell is running as an X11 client. It's receiving an X11 protocol error from the X server. These are fun because they're asynchronous, so the backtrace doesn't actually include enough diagnostic information to do anything about it. But if it really happens 50% of the time, then hopefully a developer would be able to reproduce....
He's not the only one saying that, though. See this bug comment:
"Just found out that the frequent crashes only occur in wayland. I changed to Xorg and so far, I have yet to experience a crash when I close evolution mail."
https://bugzilla.redhat.com/show_bug.cgi?id=1633365#c14
I don't think that the blocker process is suitable for this bug - but that doesn't mean that we shouldn't try to get it fixed.
Things you can do if you think a bug isn't getting enough attention:
* Make sure there's an upstream bug tracking the problem.
(https://gitlab.gnome.org/GNOME/gnome-shell/issues/375 is similar - it's also a BadWindow X error caused by a ChangeProperty request, but since nobody mentions closing windows, it's probably not the same. The debugging advice in https://gitlab.gnome.org/GNOME/gnome-shell/issues/375#note_376065 to put GDK_SYNCHRONIZE=1 in /etc/environment is useful.)
* (Politely) mail the people mentioned in the package ChangeLog (I've added a couple of relevant people to the cc:) - it can be very hard to keep on top of bugzilla mail and sort out the useful stuff.
* Mail this list (for desktop issues), and ask about it :-)
Regards, Owen
On Thu, Mar 21, 2019 at 7:35 PM John Mellor john.mellor@gmail.com wrote:
Is it possible to add fatal issue 1633365 to the list of blocker bugs for F30?
Almost everyone running F28 and F29 Gnome desktop worldwide has reported some variant of this very high severity evolution bug that easily causes a full desktop crash. I believe that its time to get this bug to block the next release.
ref: https://bugzilla.redhat.com/show_bug.cgi?id=1633365 _______________________________________________ desktop mailing list -- desktop@lists.fedoraproject.org To unsubscribe send an email to desktop-leave@lists.fedoraproject.org Fedora Code of Conduct: https://getfedora.org/code-of-conduct.html List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/desktop@lists.fedoraproject.or...
On Fri, 2019-03-22 at 09:52 -0400, Owen Taylor wrote:
I don't think that the blocker process is suitable for this bug - but that doesn't mean that we shouldn't try to get it fixed.
Things you can do if you think a bug isn't getting enough attention:
- Make sure there's an upstream bug tracking the problem.
(https://gitlab.gnome.org/GNOME/gnome-shell/issues/375 is similar - it's also a BadWindow X error caused by a ChangeProperty request, but since nobody mentions closing windows, it's probably not the same. The debugging advice in https://gitlab.gnome.org/GNOME/gnome-shell/issues/375#note_376065 to put GDK_SYNCHRONIZE=1 in /etc/environment is useful.)
Looking through the /var/log/messages attachments - these are extracts from the journal around the time of the crash, libreport provides this - about half of them have something like this right before the 'badwindow' error:
Dec 16 03:03:16 red-dwarf org.gnome.Shell.desktop[2790]: libinput error: client bug: timer event5 debounce short: offset negative (-2ms)
I'm not sure if it's significant, though.
On Thu, 2019-03-21 at 19:34 -0400, John Mellor wrote:
Is it possible to add fatal issue 1633365 to the list of blocker bugs for F30?
Almost everyone running F28 and F29 Gnome desktop worldwide has reported some variant of this very high severity evolution bug that easily causes a full desktop crash. I believe that its time to get this bug to block the next release.
FWIW, I run Evo on three systems, on F28, F29 and F30, and haven't ever seen that. (Though I don't exit it very often; perhaps you could use this as a workaround. :>)
Also note that several reporters *weren't* doing anything with Evo when it crashed...e.g. comment #52, comment #57, #59, #44, #41, #27, #25, #23.
Shell crashers can be release blockers if they're common enough, it depends on several factors. This is also quite a messy one, because the backtrace in fact just shows that GNOME crashed on an X protocol error:
Sep 26 15:10:56 paul-fedoradesktop gnome-shell[1171]: The program 'gnome-shell' received an X Window System error. This probably reflects a bug in the program. The error was 'BadWindow (invalid Window parameter)'. (Details: serial 34996 error_code 3 request_code 18 (core protocol) minor_code 0)
which means the trace isn't a lot of use, we need to know what caused that error to happen. It also means some of the dupes may not actually be the same bug at all.
I'll look into it a bit...
Anyone can propose any bug as a release blocker, there is no special requirement for this. You do it at https://qa.fedoraproject.org/blockerbugs/propose_bug .
desktop@lists.fedoraproject.org