SUMMARY When krunner is expanding due to a search, the item that happened to by under the mouse cursor by chance is executed on hitting enter. Since krunner allows commands to be executed and has plugins with potential data loss (such as a shutdown), I marked this as grave STEPS TO REPRODUCE 1. Place the mouse somewhere in the middle of the screen, upper half 2. Open krunner with a keyboard shortcut 3. Type to search 4. Hit enter OBSERVED RESULT Not the first result is chosen, but rather the one that accidentally was under the mouse cursors when krunner expanded to show search results EXPECTED RESULT Krunner should only activate that item either on mouse click, or potentially better, only do a mouse selection if there was a mouse move event. Otherwise the user most likely used the keyboard and the mouse selection is an accident. SOFTWARE/OS VERSIONS Any, really. Plasma is 6.4.4, but other versions are affected just the same.
Are you on X11? I can't reproduce this on Wayland.
(In reply to Nate Graham from comment #1) > Are you on X11? I can't reproduce this on Wayland. you are correct, this seems to indeed only happen on X11
Thanks. I'm gonna be honest there; there's little to no chance a minor X11-specific issue like this will be fixed, so keeping the bug report open it just going to provide false hope. I'd recommend transitioning to the Wayland session.
(In reply to Nate Graham from comment #3) > Thanks. I'm gonna be honest there; there's little to no chance a minor > X11-specific issue like this will be fixed, so keeping the bug report open > it just going to provide false hope. I'd recommend transitioning to the > Wayland session. Wayland is not in a usable state for me to be a daily driver yet, and I kindly ask you to leave this bug open and not just go ahead and close it already solely based on your personal views on it, thank you.
It's not my personal view; it's reality. No major Plasma developers are still using X11. Many have transitioned to using distros that don't even ship an X11 session (Like Fedora), and several are using distros that *can't* have an X11 session (Like KDE Linux). That's just the way things are, I'm afraid. So if you can't provide a patch to fix this, there is likely no chance of it getting fixed, and it's not worth keeping bug reports open for minor X11 bugs we know are likely never going to be fixed. Like I said, they provide false hope. I'd recommend submitting bug reports instead about the issues you encounter with the Wayland session that make it not usable for you.
(In reply to Nate Graham from comment #5) > It's not my personal view; it's reality. No major Plasma developers are > still using X11. Many have transitioned to using distros that don't even > ship an X11 session (Like Fedora), and several are using distros that > *can't* have an X11 session (Like KDE Linux). > > That's just the way things are, I'm afraid. > > So if you can't provide a patch to fix this, there is likely no chance of it > getting fixed, and it's not worth keeping bug reports open for minor X11 > bugs we know are likely never going to be fixed. Like I said, they provide > false hope. > > I'd recommend submitting bug reports instead about the issues you encounter > with the Wayland session that make it not usable for you. Nate, this is not the first bug of me that you close very quickly, and it wouldn't be the first that then gets re-opened and then fixed. Also while this is getting a bit off-topic: then please don't advertise X11 as still supported, but for bug fixes and no new features. Because this clearly would be a bug fix, and imho not just a small one, since this can trigger commands which result in data loss. Also, just as a PS, this bug report was reported from Fedora, which according to you does not support X11. And as far as reporting bugs for Wayland to make it usable goes: while this is always what I do hear from KDE, and that they care about making it easier for people to migrate, a good amount of people are met with "no, this is not planned", "this workflow is not supported" or "this is a design decision", which certainly doesn't help. Again: please leave it open, we will see if someone comes up with a fix, since chances are that this does also affect other components.
Can reproduce. Digging around in other widgets that I couldn't reproduce the issue in, I've found Nate actually seems to have fixed this exact issue in Kickoff a good while ago: https://invent.kde.org/plasma/plasma-desktop/-/merge_requests/542. However, the fix is written in QML, and Krunner is C++, the logic will need ported. I can't guarantee but that's what seems to be the issue, we want Krunner to "eat" the focus if invoked by shortcut. Wayland happens to sidestep the issue because of it's inherent focus behavior. PS and TL;DR, I readily use X11 every day for work, the reasons I cannot use Wayland are well-known and/or by design (VNC/remote access, window positioning). It's hard trying to help count and fix remaining X11 issues, if the reports are closed and locked.
A patch that fixes the issue will be accepted if it works and doesn't regress anything. It's your time and you can spend it however you want, but it won't change the fact that the Plasma X11 session is dying, and will be dead in a few years. Focus has shifted, resource have been re-allocated, yadda yadda yadda. Pick your favorite euphemism. Before that happens, X11-specific bug reports are going to increasingly get closed as "sorry, no resources to care about this for free" which is a normal outcome in a project with limited resources. But like I said, if someone wants to fix this, they're welcome to.
*** Bug 507367 has been marked as a duplicate of this bug. ***
*** Bug 507835 has been marked as a duplicate of this bug. ***