Bug 292091
| Summary: | [StyleBench] Make sure we don't include pending layout work in measurements. | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Emilio Cobos Álvarez (:emilio) <emilio> |
| Component: | CSS | Assignee: | Emilio Cobos Álvarez (:emilio) <emilio> |
| Status: | RESOLVED FIXED | ||
| Severity: | Normal | CC: | koivisto, webkit-bug-importer |
| Priority: | P2 | Keywords: | InRadar |
| Version: | WebKit Nightly Build | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
Emilio Cobos Álvarez (:emilio)
I'm looking into a stylebench regression in Firefox due to an scheduling
change (https://bugzil.la/1961181).
It seems that if we try to run layout / paint more often, then
stylebench improves because the test is measuring pending layout / style
work from other subtests and such.
Part of this is bug 219785 (which isn't in Firefox's tree). But it seems
worth updating layout at the beginning of the test too, to make sure
that the test measures exactly what's intended.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Emilio Cobos Álvarez (:emilio)
Pull request: https://github.com/WebKit/WebKit/pull/44550
EWS
Committed 294345@main (db0047ab0e76): <https://commits.webkit.org/294345@main>
Reviewed commits have been landed. Closing PR #44550 and removing active labels.
Radar WebKit Bug Importer
<rdar://problem/150375261>