Created attachment 105066 [details] Panel is miniscule. Users cannot use this panel because things are too small. I have a kubuntu 16.04 computer deployed and apparently after every restart after the update, the width of the left panel was reset to narrow, rendering the panel useless to the user. Since this is deployed to people with limited capability or desire to mess with settings, this is a severe bug, and I must find a solution ASAP. I was able to reproduce it in a virtual machine...
I found that new entries were being written into plasmashell.rc so I deleted all the extra ones (the last one wins apparently) and made it read only... Then plasmashell gave me a warning that the file was not writable! This is also a bug... People who lock down their settings are not experiencing a bug.
To top it off, even though the plasmashellRc was edited and made not writable the bug still occurs.
It also affects 5.8.5 if I change the screen height, the panel goes antman small.
Created attachment 105067 [details] THis is roughly how it should look I think I didn't set the windows can cover setting... that also gets lost.
Created attachment 105068 [details] change the y screen resolution and the computer is unusable
Do you still have your plasmashellrc? Are you rebooting with a different sized screen?
It is so trivial to recreate, I'm sort of offended that it was set up as striked thru waiting for info... I'm right here struggling to figure out how to solve this problem so I can let the person use the computer! I'll send you the rc
also, right clicking on the desktop now makes new note widgets... right now I'm not able to get into a konsole
Created attachment 105069 [details] there was a history line, that is completely innapropriate for being in a config This is the plasmarc.txt I have no idea why panel 1 would be there so many times... it appears that the last one wins. I tried making this readonly... and it complained There was a history line in here... which is session information... not configuration.. and I am angry that that was in here, and it has nothing to do with the configuration of my machine... I deleted that history line before I realized that I should save a copy to share with people. Why is there session data still mixed with configuration data? anyway... thickness should probably be a few hundred pixles, but it keeps on coming up really thin.
It's not that the last one wins, it's that it's storing the panel thickness setting per monitor-size. [PlasmaViews][Panel 1][Vertical899] Is the panel 1 setting when the monitor is 899 pixels high. An issue with virtual machines - but less so on real setups. Without it you end up with a different bug.
Part of this may be related to installing virtualbox additions which allowed the display to change size unexpectedly. If the size changes (like when a game is played full screen) I imagine it will change size... I'm going to test.. but assume this is it... I need to make this panel width be immutable always...
also note... some widget settings are borked, (all of my system tray icons that i disabled are back, and ones that I use i always want visible, and they are hidden) and the view mode of the panel changed from windows can cover to, windows cannot cover... visability from 2 to 0 I was able to make a config file with common vertical sizes 480 640 800 768 764 960 1024 1344 1340 1050 1080 1920 (should be others) and replicate the panel1 entry in plasmashellrc and the panel at least had the size and the viewmode back. There should be wildcards for the size... You shouldn't even worry about the desired height unless someone specifies it, so the width of the panel should not worry about the height at all. I started reading the panel.qml and saw quite a few errors from plasmashell when I resized the display, and was generally using it... I understand that there is a 5.8.6 out, and I hope it will make it to kubuntu backports soon.
I did a work around by adding panel lists in the plasmashellrc I hope it works long enough for a patch to fix this.
Related bug https://bugs.kde.org/show_bug.cgi?id=369312
*** This bug has been marked as a duplicate of bug 369312 ***