[opensuse-factory] Snapshot 0619 held for release - again i586 issues
Hello everybody, it shows that barely any developer and maintainer is using i586 systems anymore: once again, the snapshot is basically ready but fails on i586 tests (not exactly as bad as last time around, when the kernel would not even boot) This time, it's 'just' KDEs screenlocker that is crashing - See https://openqa.opensuse.org/tests/213533#step/first_boot/4 How is the feeling amongst i586 users? Shall I hold this snapshot back until the KDE Team can potentially find out what's wrong (I don't think there is currently any lead, so this can be a long wait). Please, let the i586 users get a say here: we x86_64 users won't realize it and of course 'we' would all vote to release... but that's unfair. Looking forward for your voices! Dominique
On Monday 2016-06-20 22:48, Dominique Leuenberger / DimStar wrote:
it shows that barely any developer and maintainer is using i586 systems anymore: once again, the snapshot is basically ready but fails on i586 tests (not exactly as bad as last time around, when the kernel would not even boot)
This time, it's 'just' KDEs screenlocker that is crashing
The i686 machinery I still have or are in purview of is servers which will pretty much never run KDE, and only X very little (via ssh -X).
How is the feeling amongst i586 users? Shall I hold this snapshot back until the KDE Team can potentially find out what's wrong (I don't think there is currently any lead, so this can be a long wait).
Unless it's a systematic issue -- kernel, libc or runtimes like perl as a whole and messing up the layers above it --, it almost always turns out to be some original and software-specific error. Unaligned access, shitty thread locking, you name it. As a die-hard developer, I would want to know about all bugs in my own software. But who knows where KDE's priorities are :) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Mon, 2016-06-20 at 22:48 +0200, Dominique Leuenberger / DimStar wrote:
Hello everybody,
it shows that barely any developer and maintainer is using i586 systems anymore: once again, the snapshot is basically ready but fails on i586 tests (not exactly as bad as last time around, when the kernel would not even boot)
This time, it's 'just' KDEs screenlocker that is crashing - See
For reference, this has been logged in bugzilla as well - we don't want to lose it after all. Cheers, Dominique
On Monday 20 June 2016 22:48:01 Dominique Leuenberger / DimStar wrote:
it shows that barely any developer and maintainer is using i586 systems anymore: once again, the snapshot is basically ready but fails on i586 tests (not exactly as bad as last time around, when the kernel would not even boot)
This time, it's 'just' KDEs screenlocker that is crashing - See
https://openqa.opensuse.org/tests/213533#step/first_boot/4
How is the feeling amongst i586 users? Shall I hold this snapshot back until the KDE Team can potentially find out what's wrong (I don't think there is currently any lead, so this can be a long wait).
Please, let the i586 users get a say here: we x86_64 users won't realize it and of course 'we' would all vote to release... but that's unfair.
I am actually using Plasma on i586 so it would be good to know if Plasma on i586 does not work anymore. However, I would not mind if I don't get updates for some times until someone can fix the issue. Releasing the snapshot also for i586 although it's broken might IMHO shine a bad light on openSUSE Tumbleweed as a whole even though the screensaver is actually giving an advice how a workaround could look like ;-)
On Mon, 2016-06-20 at 23:50 +0200, Oliver Kurz wrote:
I am actually using Plasma on i586 so it would be good to know if Plasma on i586 does not work anymore. However, I would not mind if I don't get updates for some times until someone can fix the issue. Releasing the snapshot also for i586 although it's broken might IMHO shine a bad light on openSUSE Tumbleweed as a whole even though the screensaver is actually giving an advice how a workaround could look like ;-
There is a mis-understanding to be sorted out here: it's all or nothing.. Tumbleweed is ONE snapshot for i586 and x86_64. Unless we start treating i586 like the other ports (ppc64le, arm), this is what we have. I have absolutely no objection to someody stepping up and doing i586 as a port - it will mean it will be completely dropped from the Tumbleweed FTP tree though... That's a completely different way of doing it (and as all things: it's not even that easy: we still need to build i586 for all packages that have a baselibs.conf and their dependencies - or we won't have the -32bit packages, which is a clear no-go). So, in short: holding back the snapshot means for ALL Tumbleweed users. Of course releasing it, Oliver can 'hold it back on his own machine' for as long as needed... and/or he can help in tracking down the bug... And now re-read above sentence but replace 'Oliver' with your own name. YOU use i586 and are thus asked to step up and help (Sorry Oliver, did really not want to single you out). Cheers, Dominique
On Tue, Jun 21, Dominique Leuenberger / DimStar wrote:
So, in short: holding back the snapshot means for ALL Tumbleweed users.
Doing so gives perhaps more time to reduce the gcc6 failures, which makes the next snapshot even more appealing. Olaf -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Dominique Leuenberger / DimStar composed on 2016-06-21 00:01 (UTC+0200): ...
So, in short: holding back the snapshot means for ALL Tumbleweed users. Of course releasing it, Oliver can 'hold it back on his own machine' for as long as needed... and/or he can help in tracking down the bug... And now re-read above sentence but replace 'Oliver' with your own name. YOU use i586 and are thus asked to step up and help (Sorry Oliver, did really not want to single you out).
TBC, there's really no stepping up non-programmer i586 TW users unfamiliar with using BS can do short of learning how to use BS, correct? IOW, the QA failure means broken packages aren't hitting the mirrors, and thus we who simply zypper up periodically wouldn't experience the screensaver breakage (assuming we have screen saving enabled in the first place, which I never do purposely)? -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Il giorno Mon, 20 Jun 2016 22:48:01 +0200
Dominique Leuenberger / DimStar
How is the feeling amongst i586 users? Shall I hold this snapshot back until the KDE Team can potentially find out what's wrong (I don't think there is currently any lead, so this can be a long wait).
As part of the team: I'd say release it, so backtraces can be obtained easier. Unfortunately I usually don't have time to check openQA myself, and even in a best case scenario, that would have to wait until the weekend. Unless someone can get a backtrace from the crashed process in openQA. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Il giorno Mon, 20 Jun 2016 22:48:01 +0200
Dominique Leuenberger / DimStar
How is the feeling amongst i586 users? Shall I hold this snapshot back until the KDE Team can potentially find out what's wrong (I don't
I can't speak for the rest of the team, but I tried, and failed to get a reproducible test case: - plasmashell crashes in my VM - running the screen locker manually does lock (doesn't crash unlike openQA) I lack sufficient time and understanding to tackle the issue in a VM, so I would propose for the snapshot to get published. If people have issues with it, run /usr/lib/libexec/kscreenlocker_greet --testing and get a backtrace. -- Luca Beltrame - KDE Forums team GPG key ID: A29D259B
Thanks Luca, On Tue, 2016-06-21 at 07:50 +0200, Luca Beltrame wrote:
Il giorno Mon, 20 Jun 2016 22:48:01 +0200 Dominique Leuenberger / DimStar
ha scritto: How is the feeling amongst i586 users? Shall I hold this snapshot back until the KDE Team can potentially find out what's wrong (I don't
I can't speak for the rest of the team, but I tried, and failed to get a reproducible test case:
In this case it might indeed be the best course then to get the snapshot out and have people report stack traces in the bug created for it. Once a fix is available, we can look at publishing it directly in the :Update channel as to not have a too long wait.
I lack sufficient time and understanding to tackle the issue in a VM, so I would propose for the snapshot to get published. If people have issues with it, run
/usr/lib/libexec/kscreenlocker_greet --testing
and get a backtrace.
I hope the #osc'16 talk about openSUSE in Numbers will have some info about i586 usage - It would be great to know the userbase of this platform a bit better. Cheers, Dominique
Il giorno Mon, 20 Jun 2016 22:48:01 +0200
Dominique Leuenberger / DimStar
Please, let the i586 users get a say here: we x86_64 users won't realize it and of course 'we' would all vote to release... but that's unfair.
A follow up: this error shown by openQA is far more serious, because it makes every application using specific features from QML crash, and only on i586/i686. A patch, which tweaks compile flags and hopefully fixes this, has been submitted as part of the Qt 5.6.1 submission to TW. We'll see if there are other solutions to "hotfix" this and make the snapshot releasable.
participants (6)
-
Dominique Leuenberger / DimStar
-
Felix Miata
-
Jan Engelhardt
-
Luca Beltrame
-
Olaf Hering
-
Oliver Kurz