the plasma-pa plasmoid sometimes do not show up in systemtray. Instead a "hole" (empty area) is shown, which is not clickable Reproducible: Sometimes Steps to Reproduce: 1. install plasma 5.4 and plasma-pa 2. start in plasma5 session Actual Results: plasma-pa not shown and not interactable Expected Results: plasma-pa should show every time after starting the session workaround: go into systemtray setting, deactive audio volume --> apply --> activate audio volume --> apply now plasmoid is shown
Had this problem too with 5.4.0 but with 5.4.1 seems to be fixed.
I still have this problem with 5.4.2
The problem somehow changed for me with frameworks 5.15 and plasma 5.4.2: now plasma-pa starts correctly and the network-manager plasmoid is showing the same problem instead.
Plasma-pa intermittently causes the same issue for me running 5.4.2 on Kubuntu 15.10. Cycling it in systray settings works as stated.
Created attachment 95159 [details] Missing Audio Volume indicator in systray Missing Audio Volume indicator after logon also on OpenSUSE Tumbleweed 20151022, using - KDE Releases Applications 15.08.2 - KDE Releases Frameworks 5.15.0 - Plasma 5.4.2 - Qt 5.5.1 Workaround as already described: System Tray Settings - General - Uncheck Extra Items -> Audio Volume, press Apply, recheck Extra Items -> Audio Volume, press Apply.
Same as in Comment 5. Empty space instead of audio icon after every reboot. KDE Plasma Version: 5.4.2 QT Version: 5.4.2 Kernel Version: 4.2.0-16-generic OS: Kubuntu 15.10 64-bit (Breeze Dark Theme) Same workaround works as described above.
This affects me too on Kubuntu 15.10 64bit. Currently using the Oxygen theme. Plasma 5.4.2 qt 4.5.2 Kernel 4.2.4
also valid using openSUSE snapshot 20151022 with QT 5.5.1
Yup affects me as well on Kubuntu 15.10 64-bits, Plasma 5.4.2 It worked like once or twice as far as I remember then got a white space instead of plasma-pa
I got the same problem on Arch Linux. An easy workaround would be: - Configure the tray manager. - Disable the audio applet then Apply. - Enable the audio applet then Apply. Even if there is an woking workaround for this, in my opinion this must be fixed for the next release.
I can confirm this issue on Archlinux too.
After spending most of the weekend debugging it i am quite convinced this is caused by an upstream bug in Qt https://bugreports.qt.io/browse/QTBUG-48870 We are trying to get this fixed but understand it may take a bit.
Fortunately, the issue is easily worked around by uninstalling plasma-pa and installing kmix. Still, the question is: even if this is a qt bug, why pushing plasma-pa in place of kmix that used to work just fine, before checking if plasma-pa could work fine? It seems impossible to me that this could have gone unnoticed by anyone who used plasma 5 for at least 2 hours in a row. As a matter of fact, it looks like almost every single serious plasma bug is in fact a bug in Qt (e.g. impossibility to work with projectors or external screens in a reliable way, this one, etc.). Still it would be very much appreciated if KDE could test changes against the versions of Qt that actually one can find in distros before pushing them in releases.
> before checking if plasma-pa could work fine? It seems impossible to me that this could have > gone unnoticed by anyone who used plasma 5 for at least 2 hours in a row. As impossible at it seems to you, it's the truth. This bug only manifests on machines with slow hard drives, since most of the developers use fast SSD drivers they did not see it. > Still it would be very much appreciated if KDE could test changes against the versions of Qt that > actually one can find in distros before pushing them in releases. Let me remind you that KDE is mostly a volunteer driven organization, all the help you can provide testing will make it better for the fellow users.
And pre-emptive, https://www.kde.org/code-of-conduct/ as a reminder to be considerate, respectful, supportive (as well as collaborative and pragmatic)
(In reply to Albert Astals Cid from comment #14) > > before checking if plasma-pa could work fine? It seems impossible to me that this could have > > gone unnoticed by anyone who used plasma 5 for at least 2 hours in a row. > As impossible at it seems to you, it's the truth. This bug only manifests on > machines with slow hard drives, since most of the developers use fast SSD > drivers they did not see it. > This bug manifests on fresh installs of Kubuntu 15.10 on two systems with a Samsung 830 and a Corsair Force LS SSD respectively. Its not limited to slow hard drives.
I am also using SSD drives but still have the issue.
I can confirm that, I am using a Lenovo X220 with SSD.
I am constantly seeing the issue on a laptop with SSD disk exactly in the same way in which i constantly see it on a desktop with a rather slow software raid built on rotating media. Makes no difference at all. The only difference that I see is that the desktop machine takes 2 minutes from the complete filling of the startup progress bar to getting an environment with all the plasmoid populated (evidently hanging on some plasmoid), while the laptop does this in a few seconds. Incidentally, my remark was referring to the fact that for almost every single bug I opened about plasma 5 either here or more frequently on ubuntu launchpad (and they were more than 20 in the first two hours of use), I found that the issue was in Qt and, often, in the specific version of Qt found in my distro. While this casts some shade on the reliability of Qt as a platform, I think that it would be very important to assure that before pushing out a new feature in plasma it is tested against the libraries that can actually be found in distros. If the feature does only work reliably on developers machines that may have some extra-new in-development version of Qt and not with the distro version of Qt, my suggestion is to leave it cooking for later. Releasing it immediately means that users will hit against bugs in Qt on their platform and attribute the issue to KDE, with a huge damage to the KDE reputation. For the issue at hand, to me this boils down to reverting to kmix (that works just fine) and keep plasma-pa on hold, until distros can catch up with a version of Qt that can support plasma-pa without issues. I thought that this could have been considered a constructive suggestion. If it is not, or if it is against the KDE code of conduct, please take my apology and feel free to ignore it.
I should have used a better working, disk speed is one of the contributing factors that makes this bug come or go away, obviously not he only one, also number of items on the systray, etc have a factor of influence on whether the bug manifests itself or not. For example, on each boot i will have the bug, if then go to a shell, do killall plasmashell; plasmashell & it will work fine. And that's basically restarting plasmashell from scratch, so yes, it is not as easy to reproduce as people that have it all the time think.
I have a machine where it does happen repeatedly. Qt model test on the proxy shows everything is fine A manual inspection of all the changes all look sensible That warning about anchors is unrelated. If I fix that this problem remains. Setting a giant cache buffer should mean all delegates are "in view" at any point. This fixes the test in the linked Qt report perfectly, but it doesn't fix our case, which implies ours is subtly different to the Qt bug report. Setting the grid size to a fixed huge size (so delegates are always in view) doesn't fix anything Nor does making sure delegates have a fixed size Swapping from GridView to Grid/Flow + Repeater causes a crash in Qt. Possible easier to debug. Possibly a different issue. Adding a shoddy timer delay before loading the applets might be enough, but I don't want to go down that route.
I'm observing the same bug but in my case it's Instant Messaging applet. Fedora 22 plasmashell 5.4.3 Qt 5.5.0
Same bug over here with Audio volume. Workaround must be applied at each start-up !
I have this same problem but with the bluetooth icon. Should I open a separate bug?
Nick, no you should not. As I already mentioned this is a bug in Qt and has been fixed for Qt 5.6 with https://codereview.qt-project.org/#/c/141620/ I suggest you guys contact your distributions and let them know about the existance of this bug and ask for a backport of the patch to the Qt they ship if you think this is critical enough.
I would not hold my breath waiting for distros to backport things or apply patches downstream E.g., the issue with hplip not being able to work on plasma 5 has been having a patch since 2014-07-29 thanks to Harald Sitter, but still, after more than 1 year the patch is not applied in (K)ubuntu... Is there any news of some distro picking up Qt 5.6 for their next release? Will (K)ubuntu 16.04 be 5.6 or 5.5.x?
This is really not the place to discuss about what KUbuntu or any other distro will or will not do.
Hi, I patched our qt5 with patch above (together with some other necessary changes) , but still have the same problem. Can you confirm it is the same issue? (Or, did you try the qt 5.6 branch?)
Yes, i did try the patch and it did seem to fix the problem for me. I tried the patch on top of my heavily patched Qt 5.4
When I clicked on the "hole" on my system, I can see some warning from plasmashell input, which says "modelData" undefined (qml file is PlasmoidItem.qml). And I tried to print out the modelIndex, and found it's -1. So this is not a visual problem but the attached model infomation is wrong. While the linked QT bug itself looks like a pure visual problem. I'm afraid that QT-BUG 48870 is not the only problem.
@Weng Xuetian also patched it here on Qt 5.5.1 but still getting the same error. So patching Qt alone does not fix it apparently.
I could always reproduce this bug doing echo 3 > /proc/sys/vm/drop_caches as root and then as my user killing and restarting plasmashell. But now I can not reproduce it anymore. Neither in my two Ubuntu Xenial systems running Plasma 5.4.3 + Qt 5.5.1 nor on my Arch Linux system running Plasma 5.5.1 + Qt 5.5.1 so can't help with fixing this anymore. If you guys are convinced the above patch does not fix it reopen the bug.
(In reply to Albert Astals Cid from comment #32) > ... > > But now I can not reproduce it anymore. Neither in my two Ubuntu Xenial > systems running Plasma 5.4.3 + Qt 5.5.1 nor on my Arch Linux system running > Plasma 5.5.1 + Qt 5.5.1 so can't help with fixing this anymore. > > If you guys are convinced the above patch does not fix it reopen the bug. Can you attach your patch for Qt 5.5.1 so I can check with mine. Maybe I missed something important. (and others too) And did you really mean Plasma 5.5.1 ? Does it add a delay or other changes that 5.5.0 maybe does not have to starting the tray or icons or plasma-pa ?
You misunderstood me, I can't reproduce the problem even without the patch.
(In reply to Albert Astals Cid from comment #34) > You misunderstood me, I can't reproduce the problem even without the patch. To be extra clear, I could before, but I can't anymore, hence the "I can't help with fixing this anymore."
(In reply to Albert Astals Cid from comment #32) > I could always reproduce this bug doing > echo 3 > /proc/sys/vm/drop_caches > as root and then as my user killing and restarting plasmashell. > > But now I can not reproduce it anymore. Neither in my two Ubuntu Xenial > systems running Plasma 5.4.3 + Qt 5.5.1 nor on my Arch Linux system running > Plasma 5.5.1 + Qt 5.5.1 so can't help with fixing this anymore. > > If you guys are convinced the above patch does not fix it reopen the bug. Personally I don't have problem on my system with 5.4.3. I tried to bisect the problem, and actually found it's a007f294b4991c6f5730e0ec8d9d20f36e31acab, which is quite unrelated to the original problem. More specifically, it's this change below. But anyway it seems to be a Qt bug since if I increase the time (0 -> 300), then I will be able to reproduce the bug will be it no matter which syntax it is used on my system. diff --git a/applets/systemtray/plugin/host.cpp b/applets/systemtray/plugin/host.cpp index 27ca227..61e8705 100644 --- a/applets/systemtray/plugin/host.cpp +++ b/applets/systemtray/plugin/host.cpp @@ -99,7 +99,7 @@ Host::Host(QObject* parent) : QObject(parent), d(new HostPrivate(this)) { - QTimer::singleShot(0, this, SLOT(init())); + QTimer::singleShot(0, this, &Host::init); } Now I workaround the bug by using "Always show all icon" option in system tray. BTW To reproduce this bug, I need to have at least 10 icons in the system tray.
(In reply to Weng Xuetian from comment #36) > ... > > diff --git a/applets/systemtray/plugin/host.cpp > b/applets/systemtray/plugin/host.cpp > index 27ca227..61e8705 100644 > --- a/applets/systemtray/plugin/host.cpp > +++ b/applets/systemtray/plugin/host.cpp > @@ -99,7 +99,7 @@ Host::Host(QObject* parent) : > QObject(parent), > d(new HostPrivate(this)) > { > - QTimer::singleShot(0, this, SLOT(init())); > + QTimer::singleShot(0, this, &Host::init); > } > You are my personal hero man ;) Yes indeed reverting that patch fixes the issue for me too. So changing from Plasma 5.5 defaults > QTimer::singleShot(0, this, &Host::init); to > QTimer::singleShot(0, this, SLOT(init())); and I get my plasma-pa icon on every boot :)
At first these look identical, but on closer inspection of QTimer: void singleShot(int msec, Qt::TimerType timerType, const QObject * receiver, const char * member) void singleShot(int msec, const QObject * receiver, PointerToMemberFunction method) behave differently when msec == 0 the first version has a shortcut path that does a QMetaObject::invokeMethod with QtQueuedConnection. The latter does not and goes via an actual timer; which possibly means it's called directly? Definitely not the true "cause" of the bug, but I'm willing to believe this has a significant impact on whether we hit the bug or not. +1 to merging this.
Git commit a0578e6e5dbd429366e6a9837af543a4c85b5a75 by David Edmundson, on behalf of Xuetian Weng. Committed on 15/12/2015 at 12:54. Pushed by davidedmundson into branch 'Plasma/5.5'. Workaround a system icon tray issue in bug QTimer::singleShot(time, QObject*, char*) QTimer::singleShot(time, QObject*, PointerToFunction) behave differently when time == 0 So the porting had an effect This patch migrates to QMetaObject::invokeMethod which is clearer There seems to be two different bug in 352055 but looks similar, the other one is fixed in Qt 5.6 according to Albert Astals Cid. M +3 -2 applets/systemtray/plugin/host.cpp http://commits.kde.org/plasma-workspace/a0578e6e5dbd429366e6a9837af543a4c85b5a75
*** Bug 356605 has been marked as a duplicate of this bug. ***
*** Bug 356555 has been marked as a duplicate of this bug. ***
*** Bug 352580 has been marked as a duplicate of this bug. ***
*** Bug 347554 has been marked as a duplicate of this bug. ***
Does this bug affect the plasma-pa icon only, or any icon in the systray ? Because it happens on my machine with plasma-pa, and it just happened with the kde-connect icon too.
It also happens with the Media Player. If both are active at boot the Media Player doesn't show, if Media Player isn't active pa doesn't show.
it affects the systray in general not just plasma-pa. Plasma PA just seemed to get hit the most often; probably has the slowest startup time or something.
*** Bug 356705 has been marked as a duplicate of this bug. ***
The problem disappeared and came back... Again with plasma-pa. My system: Archlinux Plasma 5.5.1 (on Arch, the package is plasma-workspace 5.5.1.1-1) Frameworks 5.17. I reopened... Feel free to tell me to open a new bug if needed.
Killing and starting again plasmashell solves the issue...
(In reply to Olivier Churlaud from comment #49) > Killing and starting again plasmashell solves the issue... You can also try toggling off and on plasma-pa widget in systemtray settings
*** Bug 356846 has been marked as a duplicate of this bug. ***
definitely not just plasma-pa while plasma-pa is not there on almost every reboot my ibus icon is also not shown - but much more randomly - probably on about 50% of reboots it isn't there. ibus is part of the "application status" part of sys tray so toggling that setting gets it back. I've observed this bug consistently from plasma 5.4.2 through to 5.5.1 on openSUSE Leap
Anything changed with Plasma 5.5.2?
On Arch this bug is still happening consistently across three computers all running 5.5.2. Its always the media player or volume applets that go missing. Stock everything, so no qt 5.6 git or anything.
ArchLinux, Plasmashell 5.5.2, bug is still there.
Created attachment 96559 [details] half an icon so plasma 5.5.3 is out and this bug appears to be getting worse - not better. Up until now the klipper icon in the tray has never been effeced - but now in 5.5.3 on every boot I am getting less than half a klipper icon. toggling system services in tray settings restores a full icon.
Same here on two archlinux machines with plasma 5.5.3
After updating to KDE Framework 5.18 in Manjaro system tray is gone completly crazy. Now after clean boot Klipper icon is missing (earlier only sound icon was missing), that half icon which @Mark is talking about is actually Device notifier icon and it goes away if you open hidden item by clicking on the arrow. When system boots up, if I do kquitapp plasmashell and then plasmashell, systray starts without these bugs. Anyway, I cannot beleive that systray is still broken and we are at plasma 5.5.3 version... Look at attached screenshots.
Created attachment 96648 [details] clean boot
Created attachment 96649 [details] after clicking on systray arrow
I can confirm with Nikola Schnelle's behavior with recent 5.5.3 updates which forced me to go back to 5.4.3 causing the bug to be back to the sound applet disappearing. For context i had 5.5.3 on opensuse through the KF5 + KDE repo's.
*** Bug 358319 has been marked as a duplicate of this bug. ***
I can confirm Nikola Schnelle's screenshots of the tray. It also occurs in openSUSE Tumbleweed with Frameworks 5.18 and Plasma 5.5.3. Klipper is gone, hover effect over the empty space after startup shows device notifier after clicking on the empty space there is a large text behind the symbols of the tray.
Can we have workaround implemented for this bug ("adding a shoddy timer delay before loading the applets" as David said) until fix in available in qt 5.6? Note that bug https://bugreports.qt.io/browse/QTBUG-48870 is still not fixed in qt ("in progress"). Also qt 5.6 release is delayed again and again. We have broken systray for months and we will have it broken in next months in stable product...
Git commit 13ad256cc40a657b54a236eef5c491fdf6b48ffc by Marco Martin. Committed on 03/02/2016 at 14:58. Pushed by mart into branch 'master'. reset the model on list always shown/hide change something really wrong is going on on the proxymodel updates the wrong item gets removed from the list. it may be a wrong mapping between source and dest model (doesn't seem so) or may have been some misguided attempt by QML to recycle delegates anyways resetting the model in some conditions even if expensive seems to be the only way to workaround this. Anyways this systray implementation is beyond any repair and the rewritten version won't have to rely on so many models and proxymodels Related: bug 357627 M +10 -3 applets/systemtray/plugin/tasksproxymodel.cpp M +4 -1 applets/systemtray/plugin/tasksproxymodel.h http://commits.kde.org/plasma-workspace/13ad256cc40a657b54a236eef5c491fdf6b48ffc
Git commit 42a3d8accd4e494d343954ddaa916a6c618d94f3 by Marco Martin. Committed on 04/02/2016 at 10:56. Pushed by mart into branch 'Plasma/5.5'. reset the model on list always shown/hide change something really wrong is going on on the proxymodel updates the wrong item gets removed from the list. it may be a wrong mapping between source and dest model (doesn't seem so) or may have been some misguided attempt by QML to recycle delegates anyways resetting the model in some conditions even if expensive seems to be the only way to workaround this. Anyways this systray implementation is beyond any repair and the rewritten version won't have to rely on so many models and proxymodels Related: bug 357627 M +10 -3 applets/systemtray/plugin/tasksproxymodel.cpp M +4 -1 applets/systemtray/plugin/tasksproxymodel.h http://commits.kde.org/plasma-workspace/42a3d8accd4e494d343954ddaa916a6c618d94f3
I have the same issue with audio icon disappearing, so now I went back to using kmix. However, the annoying problem is with a mysterious hidden icon which I couldn't find its originator. If I hide that icon from the System Tray entries it reappears after restarting the system. So I checked the file "plasma-org.kde.plasma.desktop-appletsrc" to see what is happening and I find that it gets a new number every time: hiddenItems=67108952,65011800 what are these numbers? and how can I find the app/service responsible for this entry? (see the attached image) Fedora 23, KDE 5.18 Qt 5.5.1
Created attachment 97062 [details] hidden icon and entry
This patch is a not proper fix, can you provide another or any ideas to resolve the issues within Qt
*** Bug 357803 has been marked as a duplicate of this bug. ***
I confirm the bug not been fixed yet.
When you say "it's not fixed yet" are you including this patch or not? https://codereview.qt-project.org/#/c/141620/ This is the one that we've said is the actual relevant fix not just attempted workarounds.
(In reply to David Edmundson from comment #73) > When you say "it's not fixed yet" are you including this patch or not? > > https://codereview.qt-project.org/#/c/141620/ > > This is the one that we've said is the actual relevant fix not just > attempted workarounds. This is not the only bug in list view unfortunately. I didn't report it to qt because I have no stable way to reproduce it. See my comment #30. I wonder replace all {begin,end}InsertRow with {begin,end}ResetModel in TaskListModel would workaround it.
The patch referenced in comment #73 (applied against qtbase-5.5.1) seems to help, at least did so the first time I started plasmashell. On my second login, plasma-pa was missing again.
It seems that this is not fixed in qt 5.6 (users on plasma irc channel confirmed this), so can you please make this top priority for fixing? Thank you.
Created attachment 97603 [details] gaps in system tray As you can see in the screenshot patch is a fail for me too. $ rpm -qa | grep qt5-qtbase qt5-qtbase-5.5.1-12.fc23.x86_64 https://bodhi.fedoraproject.org/updates/FEDORA-2016-df9f4a6560
Welcome to test this patch to see if it can workaround this issue finally. https://git.reviewboard.kde.org/r/127229/
Git commit 6a26adb50c0bfce031df55bf200ebc7f44420301 by Weng Xuetian. Committed on 01/03/2016 at 16:05. Pushed by xuetianweng into branch 'master'. Workaround by always reset model when insert and remove from model. M +6 -5 applets/systemtray/plugin/tasklistmodel.cpp http://commits.kde.org/plasma-workspace/6a26adb50c0bfce031df55bf200ebc7f44420301
Git commit 16584dc70ef9121876ded1604daf32774d1862e2 by Marco Martin, on behalf of Weng Xuetian. Committed on 01/03/2016 at 16:34. Pushed by mart into branch 'Plasma/5.5'. Workaround by always reset model when insert and remove from model. M +6 -5 applets/systemtray/plugin/tasklistmodel.cpp http://commits.kde.org/plasma-workspace/16584dc70ef9121876ded1604daf32774d1862e2
After the Plasma 5.5.5 march bug-fix update, I don't get the systray arrow to show hidden icons. To make the arrow appear I insert a usb drive to trigger the device list icon. Dropbox icon is missing too. I am using arch linux
(In reply to Raghavendra kamath from comment #81) > After the Plasma 5.5.5 march bug-fix update, I don't get the systray arrow > to show hidden icons. To make the arrow appear I insert a usb drive to > trigger the device list icon. Dropbox icon is missing too. > > I am using arch linux Same bug (Arch too). It's worst than before the patch :/
(In reply to Raghavendra kamath from comment #81) > Dropbox icon is missing too. there is a bug in Dropbox 3.14.5 (and newer) that prevents the icon loading. Nothing to do with Plasma. You should be able to get the icon back with this command dropbox stop && dbus-launch dropbox start
This patch -> https://quickgit.kde.org/?p=plasma-workspace.git&a=commit&h=42849c35780577cd9c7bd248b3086ebdfcd965df
(In reply to Mark from comment #83) > (In reply to Raghavendra kamath from comment #81) > > Dropbox icon is missing too. > > there is a bug in Dropbox 3.14.5 (and newer) that prevents the icon loading. > Nothing to do with Plasma. > You should be able to get the icon back with this command > > dropbox stop && dbus-launch dropbox start My main problem is that there is no arrow to show the hidden icons, I know dropbox has problems and it's not a major problem for me. Just to inform you that command did nothing , the dropbox icon doesn't come back, again I am stating it's not the main issue here. Because of the missing arrow we can't access notifications printers devices and other icons which are hidden.
(In reply to Raghavendra kamath from comment #81) > After the Plasma 5.5.5 march bug-fix update, I don't get the systray arrow > to show hidden icons. To make the arrow appear I insert a usb drive to > trigger the device list icon. Dropbox icon is missing too. > > I am using arch linux Comment 84 for arrow patch Let's make more clarifications, release date is simple a date, all that present in stable braches, i.e. Plasma/5.5, is for daily use and stable. source=('git://anongit.kde.org/plasma-workspace.git#branch=Plasma/5.5')
*** Bug 359711 has been marked as a duplicate of this bug. ***
Every time I get an update in Fedora 23 (updates-testing), I test out the system tray. First thing I look for is the missing audio icon, hidden icons, then the system tray configuration itself. So far in the latest update (I saw what it was KDE 5.5, I forgot the rest, I'll explain) the system tray is finally working great. No missing audio icon after restarting plasmashell 3 times, I can get to the hidden system tray icons and go into the settings all without any problem. This problem seems like it might finally be fixed once and for all? The reason I forgot is because kinfo only launched once for me without a problem. Coming here to say how great things are caused me to open it again and it won't open anymore :(
yes, things seem to be working well for me also with the update to plasma 5.5.5 on openSUSE Leap 42.1
Systemtray implementation is changed, this should be one of the problems fixed. the empty places problems has definitely been solved
*** Bug 361126 has been marked as a duplicate of this bug. ***
I waited a week, but it seems that the combination of Frameworks 5.20, Plasma 5.6, and Qt 5.6 on Arch has fixed the missing tray icon issue entirely. It is no longer occurring on three systems I've been testing for it all week, so if you are still suffering this bug try upgrading your system.
Whatever was fixed by the sni proxy, once it was incorporated into the latest plasma update, it broke again... this time worse than before... nothing is showing up in my system tray other than native KDE apps. At least before Pidgin/Java apps (though broken cause no context menu would come up) would at least SHOW UP there, now nothing.
Created attachment 101035 [details] missing systray icons I have also missing systray icons. It happens on two computers with Debian testing (plasma 4.6.5, frameworks 5.25.0). On the second computer it seemed to happen after installing hplip. See screenshot. There should be network manager, volume control and bluetooth. I can neither see the icons nor click on them. I can not even connect to wifi. On the other computer hplip is not installed. But I have only volume control and a huge gap.
This bug was closed as fixed (by a different implementation, therefore the bug status is "resolved unmaintained") in comment 91. If you still see problems, please open a new bugreport. (Feel free to mention that it looks similar to this bugreport or something like that.) General note: It's very likely that comments to a closed bug will be overlooked, especially if the closed bug already has nearly 100 comments.