Closed Bug 2086165 Opened 20 days ago Closed 4 days ago

Avoid needless flex item reflows and intrinsic-size invalidation caused by abspos and fixed-height descendants

Categories

(Core :: Layout, enhancement)

enhancement

Tracking

()

RESOLVED FIXED
162 Branch
Tracking Status
firefox162 --- fixed

People

(Reporter: mayankleoboy1, Assigned: mayankleoboy1)

References

(Blocks 1 open bug)

Details

(Keywords: perf-alert)

Attachments

(3 external files)

No description provided.

AbsoluteContainingBlock handles container-size dependencies after the containing block size is final, so these frames do not need to force additional flex item reflows.

Assignee: nobody → mayankleoboy1
Status: NEW → ASSIGNED

Skip child dependency flags when a fixed block-size supplies the percentage basis independently of the parent, retaining the existing checks for special sizing cases and flexed block sizes.

Use IntrinsicDirty::None for placeholder-only child-list changes. Include the shared mochitest comparing incremental and reconstructed layout for nested flex items with abspos and fixed-height descendants; it also logs reflow counts.

respond to the review comments on the three patches and update the patch wherever needed. Dont put any comment on bugzilla.

Flags: needinfo?(hackbot)
Flags: needinfo?(hackbot)

Emilio, do you have a specific preference for https://phabricator.services.mozilla.com/D337174#11712980 /https://phabricator.services.mozilla.com/D337174#11710573 on the approach for your whack-a-mole comment?

Flags: needinfo?(emilio)

I replied there.

Flags: needinfo?(emilio)

respond to the review comments on the three patches and update the patch wherever needed. Dont put any comment on bugzilla.

Flags: needinfo?(hackbot)
Flags: needinfo?(hackbot)
Attachment #9668948 - Attachment description: Bug 2086165 - Part 1: Avoid propagating absolute-positioned block-size dependencies to ancestors. r=#layout-reviewers → Bug 2086165 - Part 1: Avoid propagating absolute-positioned block-size dependencies to ancestors. r=emilio
Attachment #9668949 - Attachment description: Bug 2086165 - Part 2: Ignore fixed block-size descendants when checking flex item dependencies. r=#layout-reviewers → Bug 2086165 - Part 2: Ignore fixed block-size descendants when checking flex item dependencies. r?emilio
Attachment #9668950 - Attachment description: Bug 2086165 - Part 3: Avoid intrinsic-size invalidation for flex and grid placeholder changes. r=#layout-reviewers → Bug 2086165 - Part 3: Avoid intrinsic-size invalidation for non-float placeholder changes. r?emilio

Address the review comments on https://phabricator.services.mozilla.com/D337174

Flags: needinfo?(hackbot)
Flags: needinfo?(hackbot)
Attachment #9668950 - Attachment description: Bug 2086165 - Part 3: Avoid intrinsic-size invalidation for non-float placeholder changes. r?emilio → Bug 2086165 - Part 3: Avoid intrinsic-size invalidation for non-float placeholder changes. r=emilio
Pushed by mayankleoboy1@gmail.com: https://github.com/mozilla-firefox/firefox/commit/77117a9fb8e5 https://hg.mozilla.org/integration/autoland/rev/a7c59483abf3 Part 1: Avoid propagating absolute-positioned block-size dependencies to ancestors. r=layout-reviewers,emilio https://github.com/mozilla-firefox/firefox/commit/2199eb1a2f46 https://hg.mozilla.org/integration/autoland/rev/aa1fcb020504 Part 2: Ignore fixed block-size descendants when checking flex item dependencies. r=layout-reviewers,emilio,firefox-style-system-reviewers https://github.com/mozilla-firefox/firefox/commit/920eb37441ba https://hg.mozilla.org/integration/autoland/rev/3b9dcbc863d7 Part 3: Avoid intrinsic-size invalidation for non-float placeholder changes. r=layout-reviewers,emilio,firefox-style-system-reviewers
Attachment #9668949 - Attachment description: Bug 2086165 - Part 2: Ignore fixed block-size descendants when checking flex item dependencies. r?emilio → Bug 2086165 - Part 2: Ignore fixed block-size descendants when checking flex item dependencies. r=emilio

Created web-platform-tests PR https://github.com/web-platform-tests/wpt/pull/65103 for changes under testing/web-platform/tests

Status: ASSIGNED → RESOLVED
Closed: 4 days ago
Resolution: --- → FIXED
Target Milestone: --- → 162 Branch

Upstream PR merged by moz-wptsync-bot

Perfherder has detected a browsertime performance change from push 3b9dcbc863d715b37d89dd20c4c976434809c387.

No action is required from the author; this comment is provided for informational purposes only.

Improvements Test Platform Options Absolute values [old vs new] Performance Profiles
3% speedometer3 NewsSite-Nuxt/total (doc) linux2404-64-nightlyasrelease fission webrender 138.08 ms -> 133.60 ms
3% speedometer3 NewsSite-Next/total (doc) linux2404-64-nightlyasrelease fission webrender 172.62 ms -> 167.33 ms
3% speedometer3 NewsSite-Nuxt/total (doc) windows11-64-24h2-shippable fission webrender 84.52 ms -> 82.03 ms Before/After
3% speedometer3 NewsSite-Next/total (doc) windows11-64-24h2-nightlyasrelease fission webrender 109.15 ms -> 106.06 ms Before/After
3% speedometer3 NewsSite-Nuxt/total (doc) linux2404-64-shippable fission webrender 143.82 ms -> 139.75 ms
... ... ... ... ... ...
1% speedometer3 total (doc) linux2404-64-shippable fission webrender 1,498.00 ms -> 1,488.74 ms

Need Help or Information?

If you have any questions, please reach out to fbilt@mozilla.com. Alternatively, you can find help on Slack by joining #perf-help, and on Matrix you can find help by joining #perftest.

Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests.

Keywords: perf-alert
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: