KDE Plasma Users Are Still Running Into Multi-Screen Issues

[duby229]

It seems like another bug on Intel hardware, although I can't verify that. I have a multimonitor setup using an AMD card and I don't see that problem at all.

[rabcor]

Guess I lucked out on my graphics tablet then, works just fine for me on 5.6 I think, maybe I shouldn't update... :P

[callegar]

I see at least one big issue. It has to do with a weird interaction of kscreen and activities with system suspend. If you suspend a laptop with one external screen attached to it and then resume with a different external screen attached to it (very frequent for those who have "real monitors" in their offices or at home), then kscreen fails to recognize the new monitor. You need to manually kill kscreen_backend_launcher. Furthermore if you suspend a laptop with an external screen attached to it and resume without an external screen then the activity gets set up wrongly (e.g. you may get a wrong background and missing plasmoid), until you open the activities menu and re-select the activity.

But the news is that Qt at 5.6.1 or at 5.7 is evidently still broken with respect to the KDE requirements. I think that KDE in its official releases should stop relying on QT features that have still not stabilized or provide workarounds for critical behaviors, otherwise no one in the real world will ever see a stable plasma environment, because no one has development versions (or even bleeding edge versions) of QT libraries in production systems.

Needing a bleeding edge Qt also makes bug reporting very frustrating. You spend time to report an issue, wording your report in the most accurate way, possibly attempting some identification of the issue or crafting a test case. And then the report gets immediately dismissed because you are not on an advanced enough version of Qt.

I would be grateful if KDE stopped assuming that everyone can easily have bleeding edge QT on their systems. If plasma only works fine on Neon, then why having other distros package it at all? Would be good if they started testing plasma against LTS releases of QT and either provided workarounds for features that cannot be made working correctly in such a setup or delayed shipping such features in official releases.

[molekiller]

Long time KDE fan, since v2, but I think they've lost me definitively with this kind of stuff... Gnome 3 these days just work well enough for me and this kind of crap rarely happens. When they get this to work Qt6 will be out, they'll start with Plasma 6 and everything will break again, give it time...

[Fanboy80]

Well the way the settings are being stored and read are kind of strange. But once you get the settings right it works. Plasma 5.7 is definately an improvement to 5.6 but there are still some annoying bugs. You can edit ~.local/share/kscreen/<filename with random numbers> and set write protection on it as a last resort if things don't work.

[jrch2k8]

I see at least one big issue. It has to do with a weird interaction of kscreen and activities with system suspend. If you suspend a laptop with one external screen attached to it and then resume with a different external screen attached to it (very frequent for those who have "real monitors" in their offices or at home), then kscreen fails to recognize the new monitor. You need to manually kill kscreen_backend_launcher. Furthermore if you suspend a laptop with an external screen attached to it and resume without an external screen then the activity gets set up wrongly (e.g. you may get a wrong background and missing plasmoid), until you open the activities menu and re-select the activity.

But the news is that Qt at 5.6.1 or at 5.7 is evidently still broken with respect to the KDE requirements. I think that KDE in its official releases should stop relying on QT features that have still not stabilized or provide workarounds for critical behaviors, otherwise no one in the real world will ever see a stable plasma environment, because no one has development versions (or even bleeding edge versions) of QT libraries in production systems.

Needing a bleeding edge Qt also makes bug reporting very frustrating. You spend time to report an issue, wording your report in the most accurate way, possibly attempting some identification of the issue or crafting a test case. And then the report gets immediately dismissed because you are not on an advanced enough version of Qt.

I would be grateful if KDE stopped assuming that everyone can easily have bleeding edge QT on their systems. If plasma only works fine on Neon, then why having other distros package it at all? Would be good if they started testing plasma against LTS releases of QT and either provided workarounds for features that cannot be made working correctly in such a setup or delayed shipping such features in official releases.

Plasma 5 is not stable software and it should not be included on stable distros, for those plasma 4 was marked LTS to avoid this issue. So consider an rolling release distro or ask your stable distro to stop including unstable software. this was heavily announced and there is no excuse from distros to say they didn't know.

Remember plasma5 and Qt5 should be used in parallel with the same version (plasma 5.7 with Qt 5.7)until the actual re unification of the code is done, once plasma 5 is announced stable and the unification is done you can go back to the old model and support several releases of Qt with one plasma release.

Due to unification is not as simple as test with older versions because the older may actually be missing the whole piece code or even minor/major ABI/API breaking due to a post freeze refactor.

as side note seems 5.8 would be LTS and actually replace kde4 as stable note http://news.softpedia.com/news/kde-p...e-506124.shtml

P.S i have 0 mutiscreen issue on Arch with RadeonSI

[hubick]

For me this is the reason why I've switched to GNOME.

Gnome (v3) is the reason I switched to KDE. After 22 years of Linux on my desktop (you haven't lived until you've downloaded Slackware onto 50 floppy disks), I've now switched to Windows. It's really never going to be the year of Linux on the desktop, and I just want mine to work (screw freedom).

[Steffo]

Gnome (v3) is the reason I switched to KDE. After 22 years of Linux on my desktop (you haven't lived until you've downloaded Slackware onto 50 floppy disks), I've now switched to Windows. It's really never going to be the year of Linux on the desktop, and I just want mine to work (screw freedom).

Really can't understand this. For me Windows has more problems than almost any Linux Desktop.

[hiryu]

I can confirm that KDE 5.7 is only a slightly improvement for me with multi-monitor overall. There's even a few regressions.

Well before 5.7 (maybe around 5.5?), I submitted a report on a bug where closing my laptop with an external monitor connected still results in the system suspending... It was a very simply code fix that took the KDE dev around 5 minutes... but it seems the fix still didn't make it into KDE 5.7.

Not monitor related, but regression related:
Also, when I resume on my laptop, I see an issue where if I minimize or de-minimize windows after, I'll get a notice that compositing crashed, and then it restarts it. Sometimes after coming out of resume, change my refresh rate, and probably some other stuff, my desktop icon text is corrupted and I need to restart plasma:
killall plasmashell ; sleep 2 ; kstart plasmashell

Overall, 5.7 seems to work a little better than 5.6... But by a whole lot. Krunner seems to crash a bit less often.

[hubick]

Really can't understand this. For me Windows has more problems than almost any Linux Desktop.

Yeah, my laptop's desktop no longer crashes every time I turn it on, I can use the touchpad with my monitor, the HDMI connection can now display on my 4K screen @60hz (instead of 30), it folds into a tablet and the keys shut off, I can sync and update the firmware on my bike computer, use my Oculus Rift, message co-workers on Skype For Business, the list just goes on and on and on. So far, the only thing where Windows came up short of Linux was in configuring multiple static ethernet connection profiles... but I was able to get that working well enough with some Powershell scripts.