| Summary: | [regression] randomly can't rename layer in layer docker | ||
|---|---|---|---|
| Product: | [Applications] krita | Reporter: | Ming Chuan <ming> |
| Component: | * Unknown | Assignee: | Krita Bugs <krita-bugs-null> |
| Status: | REPORTED --- | ||
| Severity: | normal | CC: | halla |
| Priority: | NOR | ||
| Version First Reported In: | git master (please specify the git hash!) | ||
| Target Milestone: | --- | ||
| Platform: | Compiled Sources | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed/Implemented In: | ||
| Sentry Crash Report: | |||
|
Description
Ming Chuan
2025-10-25 13:17:18 UTC
I cannot reproduce with Qt 6.4 -- I don't have access to a newer version yet. It turns out the issue still randomly happens after reverting to older krita commit. It's more likely to be a qt 6.10 regression then. Overall qt 6.10 makes widget rendering looks more "asynchronous" and faster, so maybe some change happened upstream Yes, that sounds like a Qt issue. But should we try to patch Qt, or report upstream? Although I can't reliably repro the bug, the "QAbstractItemView::closeEditor/commitData called with an editor that does not belong to this view" warnings fortunately happens every time I edit layer names. The qt code that checks for the warning was added 2-3 years ago, but krita only starts hitting it qt 6.10 Regardless of this bug, perhaps krita's layer docker implementation should be to reviewed to confirm if krita are indeed using qt's API incorrectly as suggested by the warning message. If krita is indeed doing something incorrect then maybe fixing the warning message would also fix this bug I do think we have some weird things going on there, actually, because qt's listview editor thing is quite complex. |