Bug 312944 – [Flex] Avoid expensive percentage height walk when resolving relative position offsets

• Component: New Bugs
• Status: RESOLVED FIXED

[alan]

ssia

[alan]

Created attachment 479237 [details]
Patch

[Sammy Gill]

Comment on attachment 479237 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=479237&action=review

> Source/WebCore/rendering/RenderBoxModelObject.cpp:342
> +    if (logicalHeight.isPercentOrCalculated()) {

Do we know if we need to handle the stretch keyword here in the same way in cases where the stretch size isn't resolvable? I think from my reading of the spec it seems like they behave somewhat similar but not 100% sure. Maybe we could add a testcase for that scenario at least!

[alan]

Created attachment 479269 [details]
Patch

[alan]

> Do we know if we need to handle the stretch keyword here in the same way in
> cases where the stretch size isn't resolvable? I think from my reading of
> the spec it seems like they behave somewhat similar but not 100% sure. Maybe
> we could add a testcase for that scenario at least!
Right, I keep forgetting about 'stretch'. There's a quirks behavior difference here but otherwise should be handled as percent -Note that stretch fails on trunk -so this is a progression.

[Sammy Gill]

nice, thanks! I just remembered about aspect-ratio and we might need to add that too I think.

[alan]

Created attachment 479274 [details]
[fast-cq]Patch

[EWS]

Committed 311908@main (c7c99b8fbe3e): <https://commits.webkit.org/311908@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 479274 [details].

[Radar WebKit Bug Importer]

<rdar://problem/175466302>

[Sammy Gill]

Submitted web-platform-tests pull request: https://github.com/web-platform-tests/wpt/pull/59548