ComputePostFilterExtents and nsCSSFilterInstance::SetBounds are malloc heavy
Categories
(Core :: SVG, enhancement, P3)
Tracking
()
Tracking | Status | |
---|---|---|
firefox111 | --- | fixed |
People
(Reporter: smaug, Assigned: longsonr)
References
Details
Attachments
(1 file)
Testcase is to move mouse over https://kangax.github.io/compat-table/esnext/
See also https://bugzilla.mozilla.org/show_bug.cgi?id=1520294
https://perfht.ml/2FBxTUe
and
https://perf-html.io/public/5efdb443905498eaba6b8479e64ab16fc7601712/calltree/?globalTrackOrder=0-1-2&hiddenGlobalTracks=1&hiddenLocalTracksByPid=3149-0-1-2-3-4~3336-0-1~3287-0-1&invertCallstack&localTrackOrderByPid=3149-0-1-2-3-4~3336-0-1~3287-0-1~&search=ComputePost&thread=9&v=3
Reporter | ||
Comment 1•6 years ago
|
||
AutoTArray should work here, but I'm not sure about the best size for it.
Assignee | ||
Comment 2•6 years ago
|
||
Me neither, it's the filter chain size which is unbounded in markup. Maybe guess at 8 to start, most filter chains aren't that big.
Assignee | ||
Comment 3•6 years ago
•
|
||
For CSS property filters the most common chain size might even be 1, though not for ComputePostFilterExtents as that does SVG filters too I believe.
![]() |
||
Comment 4•6 years ago
|
||
(In reply to Robert Longson [:longsonr] from comment #2)
Me neither, it's the filter chain size which is unbounded in markup. Maybe guess at 8 to start, most filter chains aren't that big.
That seems about right to me too.
Updated•2 years ago
|
Assignee | ||
Comment 5•2 years ago
|
||
Updated•2 years ago
|
Comment 7•2 years ago
|
||
bugherder |
Description
•