| Summary: | Printers widget think printing is still occurring even though it has completed | ||
|---|---|---|---|
| Product: | [Plasma] plasmashell | Reporter: | John Veness <john.kde> |
| Component: | Printers widget | Assignee: | Plasma Bugs List <plasma-bugs-null> |
| Status: | CONFIRMED --- | ||
| Severity: | normal | CC: | martin.schnitkemper, noeerover |
| Priority: | NOR | ||
| Version First Reported In: | 6.5.2 | ||
| Target Milestone: | 1.0 | ||
| Platform: | openSUSE | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed/Implemented In: | ||
| Sentry Crash Report: | |||
|
Description
John Veness
2025-11-21 16:13:52 UTC
I can confirm this behavior. After printing, the icon remains in the system tray. When you open it, the printer is shown as "idling," and the print queue displays it as "completed." However, the last job is still displayed with the "Cancel" button, even though it's already finished. The "Cancel" button doesn't do anything because there's nothing to cancel. Only after logging out and back in to the desktop does the icon disappear. Operating System: Arch Linux KDE Plasma Version: 6.5.3 KDE Frameworks Version: 6.20.0 Qt Version: 6.10.1 Kernel Version: 6.17.9-arch1-1 (64-bit) Graphics Platform: X11 It appears we're not getting notification that the job queue is cleared, but I can't reproduce this in 6.5 or master. Any chance you could try this out with Plasma 6.6? Also, what version of CUPS are you running? I can't test Plasma 6.6 right now, but I am running CUPS 2.4.14. Alright, I have found a scenario this can happen and it appears to be related to a CUPS dbus-notifier lock that is not released or is "stuck". Plasma version doesn't matter, but if for some reason, a dbus-notifier get stuck (hangs), we stop getting notified of CUPS events, like for example, a print job finishing. Even restarting the plasma shell has no effect. Short of a complete logout, the following seems to put CUPS back into a working state: Logout of plasma Switch to a TTY login systemctl stop cups rm /etc/cups/subscriptions.* rm -rf /var/cache/cups systemctl restart cups logout Switch back to SDDM, start the plasma session As to why this happens, I'm not really sure. It's possible a crashing process that holds a notification subscription would prevent a clean "unlock" of the notifier, but I see no evidence that is happening. Plasma will generally have two processes running that "listen" for CUPS events, one is the marker level checker (kded process) and the other is the plasmoid, (plasmashell process). Both seem to operate normally and react properly to signals that are received and they log no errors or warnings in this scenario. I should follow that up with this: If you see plasmashell crash or a kded crash, then that could possibly cause this behavior because both will attempt to restart and continue on. |