Corner buttons in non-maximized windows no longer easily clickable
Categories
(Core :: Widget: Gtk, defect, P3)
Tracking
()
People
(Reporter: hjalte3, Unassigned)
References
(Blocks 1 open bug)
Details
(Whiteboard: gtk-severity-sweep-2026-10)
Attachments
(2 files)
User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:149.0) Gecko/20100101 Firefox/149.0
Steps to reproduce:
First, I'll admit I have a niche usecase. I use Firefox on niri, a tiling window manager. Windows are automatically placed side-by-side in a grid, without overlapping. I have gaps set to 0, so windows are placed directly against the screen edges. This allows the benefits of Fitt's law to apply to all the buttons in the Firefox navigation bar and sidebar, which is quite nice. Because there are no gaps, this can be true for any window depending on its exact position within the grid, whether "maximized" or not. Actually, Niri does not normally maximize windows in the traditional sense, instead even windows which fill the whole screen remain in an "unmaximized" state, and the rounded top corners remain. Despite this, it has been possible to press the buttons on the left and right ends of the navigation bar (for me, the sidebar toggle on the left, and the window close button on the right) by placing my cursor in the corresponding screen corner.
Actual results:
On a recent nightly build there has been (from my perspective) a regression where I can no longer press the buttons on the left and right ends of the navigation bar easily anymore by placing my mouse cursor in the screen edges. Mozregression points me to Bug 2005854 as the cause.
Expected results:
I would not have expected this change in behavior, especially from what appears to be a theme-related change. Being able to click the buttons in the corners of the window easily seems like a positive in terms of usability, although I guess it could be considered unexpected due to the visually rounded corners.
Comment 1•9 months ago
|
||
The Bugbug bot thinks this bug should belong to the 'Core::Widget: Gtk' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.
Comment 2•9 months ago
|
||
Can you attach screenshot to show what do you mean? Maybe show previous and recent states?
Thanks.
Updated•9 months ago
|
By the way, this can also be seen on KDE Plasma with the KZones kwin script.
I noticed that the built-in snapping/tiling in Gnome, KDE and Cosmic make all GTK windows like Firefox become square-cornered as if they "think" they are maximized when they are snapped/tiled. More or less the same thing seems to happen on Hyprland from what I can tell. So I guess niri may be the odd compositor out by default. Actually, I just found the configuration option prefer-no-csd which makes niri behave similarly (with some slightly annoying side effects). So I guess a workaround exists, at least for me.
Actually, I found that the better (more targeted) niri option to work around this issue is the window rule tiled-state true.
This still seems to be an unintended consequence/regression of bug 2005854? It's pretty minor though, especially when workarounds exist, so I can understand if this ends up as WONTFIX.
Updated•1 month ago
|
Description
•