SUMMARY *** The current behavior, while intentional, deviates from the standard used by Latte, DashToDock and MacOS. *** STEPS TO REPRODUCE 1. Use the "Floating Panel" and "center aligned" options for the main tray, leaving some room on either side. 2. Move cursor to screen edge where "empty" space is 3. Grope around until you find the panel. OBSERVED RESULT Panel does not un-hide unless the cursor hits the area where the panel is. Video: https://drive.google.com/file/d/1m8AhEK9JtxFNq1RyV9bMym7WKN4R-LFd/view?usp=drivesdk EXPECTED RESULT Panels should un-hide when cursor hits anywhere on the screen edge, unless otherwise specified, e.g. in the case of a miniature clock or media player plasmoid on the side. Current behavior doesn't make sense for a main panel, and deviates from the behavior of other popular panels like Latte, DashToDock and MacOS. SOFTWARE/OS VERSIONS SOFTWARE/OS VERSIONS Operating System: Kubuntu 22.10 KDE Plasma Version: 5.26.5 KDE Frameworks Version: 5.101.0 Qt Version: 5.15.6 Kernel Version: 5.19.0.30 Graphics Platform: X11 and Wayland ADDITIONAL INFORMATION Bonus points for a configurable "Intelligent Auto-hide" ability, where the user can select whether to show the panel unless a window takes the space etc.
You said, "Panel does not un-hide", but you didn't mention configuring the panel to be auto-hide. Can you confirm that the panel is set up to be auto-hide, and that whether it's set to "Floating Panel" mode or not is irrelevant to this behavior?
(In reply to Nate Graham from comment #1) > You said, "Panel does not un-hide", but you didn't mention configuring the > panel to be auto-hide. Can you confirm that the panel is set up to be > auto-hide, and that whether it's set to "Floating Panel" mode or not is > irrelevant to this behavior? Correct, it's set to automatically hide. This report is the cleaned up, wishlist version of another report, as requested. https://bugs.kde.org/show_bug.cgi?id=464162 Let me know if this is clear.
(In reply to tim from comment #2) > (In reply to Nate Graham from comment #1) > > You said, "Panel does not un-hide", but you didn't mention configuring the > > panel to be auto-hide. Can you confirm that the panel is set up to be > > auto-hide, and that whether it's set to "Floating Panel" mode or not is > > irrelevant to this behavior? > > Correct, it's set to automatically hide. This report is the cleaned up, > wishlist version of another report, as requested. > > > https://bugs.kde.org/show_bug.cgi?id=464162 > > > Let me know if this is clear. Ugh, I really wish there were a way to edit these. Yes, I definitely should have mentioned that. My bad.
Thanks. Re-titling for clarity.
I think this makes sense -- currently having both screen edge actions and a auto-hiding panel on the same edge is broken anyways. One concern is that this might degrade the ability to interact with other UI elements near screen edges, like scroll bars on the right. Another question is whether we want to support multiple auto-hide panels on the same edge.
(In reply to fanzhuyifan from comment #5) > One concern is that this might degrade the ability to interact with other UI > elements near screen edges, like scroll bars on the right. It's already an issue, so at least that wouldn't be a regression. > Another question is whether we want to support multiple auto-hide panels on > the same edge. That's a good point. My first thought is that if there's only one auto-hide panel, make its activation zone span the entire screen edge. And if there's more than one, don't, and give each one an activation zone linked to its shape.
(In reply to Nate Graham from comment #6) > (In reply to fanzhuyifan from comment #5) > > One concern is that this might degrade the ability to interact with other UI > > elements near screen edges, like scroll bars on the right. > It's already an issue, so at least that wouldn't be a regression. I think this would be a regression if such an auto-hide panel has small length. Currently it only interferes with the part of the edge corresponding to its shape, but with this change it will interfere with interactions on the whole edge.
True. I wonder how often people put a single small auto-hide panel on a screen edge though.
*** Bug 484589 has been marked as a duplicate of this bug. ***
I understand there hasn't been much activity on this, but it would be really nice to have this implemented, as with a smaller panel it is quite easy to miss the trigger area. The entire screen edge trigger would be amazing to have, even as some option instead of default behaviour.
A complication is what happens when someone has multiple auto-hide panels on the same screen edge? It's possible and valid.
(In reply to Nate Graham from comment #11) > A complication is what happens when someone has multiple auto-hide panels on > the same screen edge? It's possible and valid. I don't have enough knowledge to comment if these can implemented, but theoretically I can think of two solutions. 1. There can be a toggle in panel settings to bring up all the panels at once, on that particular screen edge where panel resides, while default is the current behaviour - each panel has its own trigger area. 2. Have a way to set the trigger area in panel settings with respect to panel length, 100% being the default. Increasing it significantly can help reach the state where pushing against just the screen edge will trigger it.
> My first thought is that if there's only one auto-hide panel, > make its activation zone span the entire screen edge. > And if there's more than one, don't, and give each one > an activation zone linked to its shape. This can be a clean solution as well without adding any more options/complexity.
Yeah, that sounds like it could work.
I discussed this with Vlad from KWIN and I'm currently closing this as intentional, though I am ready to discuss this again if there's *significant* interest from people. This would require changes in the protocol between Plasma and Kwin, it would probably be quite annoying on ultra-wide monitors (it would either have to be an option, which I really don't want, or only happen on certain sizes of monitors/panels, which would be arbitrary), plus other technical reasons ("kwin does not hide the panel after the pointer leaves it; you'd need to rework the entire mechanics behind panel autohiding so plasmashell doesn't need to take care of enter/leave events to avoid the panel getting stuck if you accidentally bump into a screen edge"). Right now, this does not seem to be worth the technical burden to implement and maintain it.
*** Bug 499833 has been marked as a duplicate of this bug. ***
*** Bug 509824 has been marked as a duplicate of this bug. ***