https://bugs.kde.org/show_bug.cgi?id=524031
--- Comment #5 from Adam <[email protected]> --- Test to separate the window from the process, done while the failure was active: xdotool windowunmap 0x1200008 -> pointer input restored immediately xdotool windowmap 0x1200008 -> input blocked again The process was left running throughout, so it is the mapped window that blocks input, not anything about the process being killed or the restack that follows it. Note that xdotool windowmove on this window is a no-op; the geometry stays 3840x2132+0+28 (the output work area) regardless, presumably clamped by KWin. Unmap is what works. Frequency has increased sharply: it now recurs every few hours rather than every day or two. Seven occurrences so far, all cleared by removing the window: killing the bridge six times, and the unmap test above once. -- You are receiving this mail because: You are watching all bug changes.
