Closed dbazile closed 3 years ago
Thanks a lot for the highly detailed report! I now added a test-case for this and fixed the underlying issue in: 250529d15cc80fefaae57416e1108f692a424c24
I will prep a release sometime this week. If the issue remains with this fix (it will be part of the v28
release), please let me know!
Thanks a lot!
Awesome, thanks! I've set a reminder to test this again and close this issue as soon as v28 makes its way into Fedora 33.
dbus-broker v28 made it into Fedora 33's update stream a few days ago. Upgraded, tested and verified that the session crash is fixed. :+1:
Cool, thanks a lot!
If I use a specific duplicate dbus monitoring match expressions, the entire GNOME session is killed as soon as I exit the script, kicking me out to the login screen. At one point during testing last night it even locked up the system and I had to do a hard power off/on to get back in.
It took a while to narrow down what specifically was causing the crash but I found that if there is a duplicate match that contains a
sender=...
clause, I can consistently reproduce the crash.I originally posted this at dbus/dbus-python but was redirected here.
It does look like this might be the same issue as #245 but my currently-installed dbus-broker is v27 which seems to have had the fix for #245 incorporated.
Versions
Fedora 33.
Steps to reproduce
journalctl output