Bug 490005 - plasmashell segfaults at init from git f7ee03d065b4e293746248f749a7965c4321b1cb (bisected)
Summary: plasmashell segfaults at init from git f7ee03d065b4e293746248f749a7965c4321b1...
Status: RESOLVED DUPLICATE of bug 489940
Alias: None
Product: plasmashell
Classification: Plasma
Component: Panel (show other bugs)
Version: master
Platform: Gentoo Packages Linux
: NOR crash
Target Milestone: 1.0
Assignee: Plasma Bugs List
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2024-07-10 07:38 UTC by Duncan
Modified: 2024-07-11 05:23 UTC (History)
1 user (show)

See Also:
Latest Commit:
Version Fixed In:
Sentry Crash Report:


Attachments
gdb.txt run and backtrace log (41.41 KB, text/plain)
2024-07-10 07:38 UTC, Duncan
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Duncan 2024-07-10 07:38:35 UTC
Created attachment 171523 [details]
gdb.txt run and backtrace log

plasma-workspace git f7ee03d065b4e293746248f749a7965c4321b1cb "Refactor floating panels" is the culprit -- the immediately previous e818368c22d8af554f9173fda2049439f73ab71a was both my previously working HEAD and is reconfirmed to run without segfault after a rebuild.

I'm on wayland (no X but xwayland installed), running live-git frameworks/plasma/most-gear, updated yesterday, with the segfault bisecting to the above specific commit.

100% Reproducing bisecter:
Attempted to start kde/plasma from the framebuffer login (no *DM gui login if it matters), plasmashell crashed but I got the rest of a usable session so could attempt a relaunch using krunner/konsole/whatever, only to have plasmashell crash again.  I then ultimately did the bisect and unstripped-ggdb rebuilds and ran gdb from konsole.

I should mention that I have multi-monitors; it's possible the segfault wouldn't trigger on single-monitor.

The platform is Gentoo Linux ~amd64 running the live-git kde packages from the gentoo/kde overlay.  qt 6.7.2, frameworks reporting 6.4.0 but of course it's live-git.  plasmashell --version (run in konsole) reports 6.1.80, but of course live-git as well.  I'm building with gcc-14.1.1-p20240622 p2.

I am by /no/ means a debugging/gdb expert but I rebuilt plasma-workspace (and thus plasmashell) unstripped, -ggdb, then ran gdb plasmashell, enabled logging, ran it for the segfault, did a thread apply all backtrace, detached and quit.  I'm attaching that log; hope it helps.
Comment 1 Duncan 2024-07-10 07:45:38 UTC
Just in case a fix crossed with this report...my current (still broken) plasma-workspace head: 5478b34eb8c21ce5e4d352f47507042b2d0c3e4a
Comment 2 Duncan 2024-07-11 05:23:41 UTC
Duplicate of now-fixed bug #489940

*** This bug has been marked as a duplicate of bug 489940 ***