Created attachment 133005 [details] The backtrace of the crash STEPS TO REPRODUCE 1. Try to open the Processes tab OBSERVED RESULT Crash. EXPECTED RESULT No crash. SOFTWARE/OS VERSIONS (available in About System) KDE Plasma Version: git master KDE Frameworks Version: git master Qt Version: 5.15 (from source code) plasma-systemmonitor: 098b91556901ccb611028b3a9156de61295b664c
I also see lots of qt.core.qabstractitemmodel.checkindex: Index QModelIndex(-1,-1,0x0,QObject(0x0)) is not valid (expected valid) qt.core.qabstractitemmodel.checkindex: Index QModelIndex(-1,-1,0x0,QObject(0x0)) is not valid (expected valid) qt.core.qabstractitemmodel.checkindex: Index QModelIndex(-1,-1,0x0,QObject(0x0)) is not valid (expected valid) qt.core.qabstractitemmodel.checkindex: Index QModelIndex(-1,-1,0x0,QObject(0x0)) is not valid (expected valid) qt.core.qabstractitemmodel.checkindex: Index QModelIndex(-1,-1,0x0,QObject(0x0)) is not valid (expected valid) qt.core.qabstractitemmodel.checkindex: Index QModelIndex(-1,-1,0x0,QObject(0x0)) is not valid (expected valid) qt.core.qabstractitemmodel.checkindex: Index QModelIndex(-1,-1,0x0,QObject(0x0)) is not valid (expected valid) qt.core.qabstractitemmodel.checkindex: Index QModelIndex(-1,-1,0x0,QObject(0x0)) is not valid (expected valid) in the terminal
Thanks for the report Vlad I am on ksysguard from git master as well and I can switch beteen Process Table and System Load tabs without any crash. Have you made any config changes that might assist us in replicating the crash?
Apologies I didn't realise plasma-systemmonitor was a separate application but I cannot replicate it in there either.
Created attachment 133010 [details] My config file
Created attachment 133020 [details] terminal output
FYI Justin, if there's a crash backtrace, you generally don't need to do the NEEDSINFO WAITINGFORINFO dance since the backtrace itself is evidence enough that there's a real issue. Or at least, wait a year or so just in case it got fixed. :) Also Vlad is one of the KWin maintainers so his bug reports are probably going to be real. :)
(In reply to Nate Graham from comment #6) > FYI Justin, if there's a crash backtrace, you generally don't need to do the > NEEDSINFO WAITINGFORINFO dance since the backtrace itself is evidence enough > that there's a real issue. Or at least, wait a year or so just in case it > got fixed. :) > > Also Vlad is one of the KWin maintainers so his bug reports are probably > going to be real. :) All good, I try to check the reporter every time but this one slipped through. Have set to confirmed as there is a backtrace.
A possibly relevant merge request was started @ https://invent.kde.org/plasma/plasma-systemmonitor/-/merge_requests/65
Git commit 153ce2f197085be42cf1488359a5ca4fcae1e3cf by David Edmundson. Committed on 11/11/2020 at 10:40. Pushed by davidedmundson into branch 'master'. Fix filter invalidation When we have matched columns we change what filterAcceptsRow passes. This means we need to trigger re-evaluation of filterAcceptsRow not a layout change. A layout change implies movement, but not a change in which rows exist. Without this KDescendantsProxyModel gets out of sync as it correctly doesn't refresh indexes on a layout change. M +1 -1 src/table/ProcessSortFilterModel.cpp https://invent.kde.org/plasma/plasma-systemmonitor/commit/153ce2f197085be42cf1488359a5ca4fcae1e3cf