https://bugs.kde.org/show_bug.cgi?id=492064
--- Comment #3 from Adam Fontenot <adam.m.fontenot+...@gmail.com> --- (In reply to Waqar Ahmed from comment #1) > > The popup should only be closed on a click or when the window is actually > > focused, e.g. by moving the cursor to another window and then back. > > Yeah, not sure that is a good idea. But perhaps we should increase the > timeout for hiding the popup a bit. One thought I had was that you could set a variable timeout - if the cursor is within a certain small distance of the box, only hide it after 1s or 1.5s. This would help deal with the fact that the margins on the box are really tight and it's very easy to go slightly outside it. Your change of 100ms to 300ms will help with that, but it feels more appropriate in the case when the user has *definitely* left the box, not when the cursor ventures 10 pixels outside it. Also, I think it would be a *major* QoL improvement to check whether the cursor is hovering over the *same* symbol as triggered the popup box, both at the beginning of the hover (when we start the timer) and at the end (when we close the box). If the user has their finger on a trackpad and just barely moved it after activating the popup, Kate LSP will close and then immediately reopen the box! By checking whether the same symbol as before is the hover target, we can avoid that. It's the little details that make the difference. When I wrote this bug report I tested it without hover to focus, and I felt that the behavior was much better - but trying it now I'm not sure at all that that's the case. So the whole premise of this bug report (that hover to focus is the source of the problem) may be a mistake, and in that case of course my suggestion of changing the behavior to only deactivate the box when leaving the window would be incorrect! -- You are receiving this mail because: You are watching all bug changes.
