Rawhide running very poorly for KDE

I upgraded my Fedora 41 to Rawhide and was surprised that fundamental things simply don’t work:

A) KDE + Wayland doesn’t work at all, there’s a long pause and it comes back to the SDDM. Looking into coredumpctl shows plasmashell is crashing

B) KDE + Xorg works but it very slow to login. In the Xorg/KDE desktop these don’t work:

  1. clipboard “clear history” doesn’t clear the clipboard
  2. Desktop can not be locked

Is this typical of Rawhide now? I’ve been running on Xorg/KDE for a while and this is a sample list of core dumps shown by ‘coredumpctl list’

Sat 2025-01-18 21:00:14 EST 11662 1000 1000 SIGABRT present  /usr/bin/gmenudbusmenuproxy 467.9K
Sat 2025-01-18 21:00:14 EST 11699 1000 1000 SIGABRT present  /usr/libexec/drkonqi        858.1K
Sat 2025-01-18 21:00:14 EST 11644 1000 1000 SIGABRT present  /usr/bin/kwin_x11             1.2M
Sat 2025-01-18 21:00:14 EST 11803 1000 1000 SIGABRT present  /usr/bin/kaccess            595.2K
Sat 2025-01-18 21:00:14 EST 11904 1000 1000 SIGABRT present  /usr/libexec/drkonqi        858.7K
Sat 2025-01-18 21:00:14 EST 11885 1000 1000 SIGABRT present  /usr/bin/kwin_x11             1.2M
Sat 2025-01-18 21:00:15 EST 12028 1000 1000 SIGABRT present  /usr/libexec/drkonqi        859.3K

Rawhide is not a tested release, bugs may randomly occur

See Rawhide :: Fedora Docs

My question is these processes are all immediately dying at login time preventing login, these aren’t slow edge condition style core dumps. Why are these being pushed out even on rawhide like that? I would think these critical processes instantly dying would show up with even the testing automation rawhide gets

They are pushed out for you to test.

May be a hardware issue.
I have rawhide + Kde on my laptop (intel +nvidia) and on my desktop (RTX 4070 sti) and I have no problem.
I have been using only rawhide for at least 5 years. The few problems I have had concern some critical package updates like glibc. When that happens I download the old versions on fedora koji and I do a downgrade.

FYI, critical-path rawhide builds all go through a gating CI that tests a bunch of things. It can of course miss things that are hardware specific or the like, but in general it works pretty well.

SIGABRTS are not something thats expected at all.

I’d suggest filing a bug (to start with) on plasmashell so developers can help track it down. If you want something more interactive, the #kde:fedoraproject.org matrix room may be able to help isolate whats going on.

FWIW, kde (wayland) runs fine here in rawhide. No particular problems.

1 Like

I mean they shouldn’t be pushed out so fundamentally broken they instantly core dump. Isn’t there’s a test framework that checks for that?
Anyway, the update I did on 1/21 seem to fix it, which was weird because the packages updated had nothing to do with plasmashell

@marko3209 as already mentioned, rawhide goes through automatic testing which should detect things like that.

As this is pretty technical, if the issues persist open a bug report on plasmashell. Attach a log from journalctl -p err to get better info.