ResizeObserver keeping the refresh driver ticking until the page is reloaded
Categories
(Core :: Layout, defect)
Tracking
()
Tracking | Status | |
---|---|---|
firefox99 | --- | fixed |
People
(Reporter: florian, Assigned: smaug)
References
(Blocks 1 open bug)
Details
Attachments
(3 files)
I don't have steps to reproduce, but I have a profile showing how this started. I reloaded the https://sql.telemetry.mozilla.org/dashboard/carbon-footprint-desktop-indicators page, and somehow vsync remained enabled permanently. I had the profiler running in marker-only mode, so I captured a profile: https://share.firefox.dev/3JCVKR8
Here's another profile of the same process with stack sampling: https://share.firefox.dev/3LUqC1N
This behavior stopped when I reloaded the tab, and I haven't found steps to reproduce.
Markus, any idea about how this could be made actionable?
Assignee | ||
Comment 1•3 years ago
|
||
ResizeObserver is more a layout thingie. -> layout.
Assignee | ||
Comment 2•3 years ago
|
||
https://searchfox.org/mozilla-central/rev/68a5327697ec43aa55b458c504c4b313c9c80528/dom/base/ResizeObserverController.cpp#97,99,139 is at least suspicious. I couldn't immediately see how the RefreshDriver is stopped in that case
Reporter | ||
Updated•3 years ago
|
Assignee | ||
Comment 3•3 years ago
|
||
Assignee | ||
Comment 4•3 years ago
|
||
Updated•3 years ago
|
Updated•3 years ago
|
Comment 6•3 years ago
|
||
bugherder |
Comment 7•3 years ago
|
||
I've got a regression test here that seems to fail locally without the patch, using the API from bug 1756269.
(I'm currently rebuilding off of up-to-date mozilla-central to be sure it passes with the patch, and then will request review.)
Here's a Try build with the test and without the fix, to validate that it fails i.e. that it's sensitive enough to detect a regression:
https://treeherder.mozilla.org/#/jobs?repo=try&revision=c02d10dfa6816732e4341a0fd2b79ff6d44c4a49
Comment 8•3 years ago
|
||
Comment 10•3 years ago
|
||
Backed out 2 changesets (Bug 1756118, Bug 1756269) for causing linting failures.
Backout link
Push with failures
Failure Log
Comment 11•3 years ago
|
||
I fixed the linting failure & am re-landing, as noted in bug 1756269 comment 7.
Comment 12•3 years ago
|
||
Comment 13•3 years ago
|
||
bugherder |
Description
•