| Summary: | MagicLampEffect triggers empty subquad assert from makeGrid -> makeSubQuad | ||
|---|---|---|---|
| Product: | [Plasma] kwin | Reporter: | Serhiy Zahoriya <serhiy.int> |
| Component: | effects-various | Assignee: | KWin default assignee <kwin-bugs-null> |
| Status: | RESOLVED FIXED | ||
| Severity: | crash | CC: | vlad.zahorodnii |
| Priority: | NOR | Keywords: | drkonqi |
| Version: | 5.4.2 | ||
| Target Milestone: | --- | ||
| Platform: | Ubuntu | ||
| OS: | Linux | ||
| See Also: | https://bugs.kde.org/show_bug.cgi?id=338972 | ||
| Latest Commit: | Version Fixed In: | ||
| Sentry Crash Report: | |||
|
Description
Serhiy Zahoriya
2015-10-28 06:35:41 UTC
Likely similar to https://git.reviewboard.kde.org/r/125131/ - just another grid creating function.... I wasn't able to cause this (the reason would likely be an empty quad in the list, what should not happen itfp) - can you test a patch?
diff --git a/libkwineffects/kwineffects.cpp b/libkwineffects/kwineffects.cpp
index 33e295c..57cdf8a 100644
--- a/libkwineffects/kwineffects.cpp
+++ b/libkwineffects/kwineffects.cpp
@@ -1082,6 +1082,10 @@ WindowQuadList WindowQuadList::makeGrid(int maxQuadSize) const
const double quadRight = quad.right();
const double quadTop = quad.top();
const double quadBottom = quad.bottom();
+ if (quadTop >= quadBottom || quadLeft >= quadRight) {
+ qDebug() << "D'OHHHH!" << quadTop << quadLeft << quadBottom << quadRight;
+ continue;
+ }
// Compute the top-left corner of the first intersecting grid cell
const double xBegin = left + qFloor((quadLeft - left) / maxQuadSize) * maxQuadSize;
------
This should prevent the assert bug get you some D'OHHHH line(s).
I can't reproduce it on another machine either (Having this problem on a laptop with intel video, can't reproduce on desktop with nvidia, both nouveau and official drivers). I've tried to compile KDE from sources, but anongit.kde.org is down right now and there is also a warning that it might not work in Kubuntu. Also the compiling wiki page is outdated at least for Kubuntu. WIll try again later. Martin, is this bug still relevant? I think https://phabricator.kde.org/D11015 fixed this bug too. Let's assume it's fixed. |