Bug 310264 – LCP in shadow DOM is not reported

• Component: Layout and Rendering
• Status: REOPENED

Created attachment 478720 [details]
you will need to install playwright test for this

Both chrome and firefox report LCP coming from shadow dom but webkit doesn't.
Given a scenario if image is loaded from shadow dom and is the best LCP candidate, chrome and firefox will report it as an lcp where webkit will ignore it and take only candidates from regular DOM.

I am not really sure if webkit is more behaving to spec or not.
It could be that firefox aligned to Chrome and webkit decided to follow the spec more closely.

Relevant links:

  • LCP spec: https://w3c.github.io/largest-contentful-paint/ (section 4.2)
  • Paint Timing spec: https://w3c.github.io/paint-timing/ (section 5.4.1 "Exposed for paint timing")


The LCP spec section 4.2 step 5.2 says 
If imageElement is not exposed for paint timing, given document, continue.

in Paint timing spec seciton 5.4.1 step 4
If element’s root is not equal to document, return false.

I am not sure how to interpret it.

I have attached a test script you can run via node
It uses @playwright/test in order to run all 3 browsers.

[Simon Fraser (smfr)]

I thought there was a web platform test for this that we passed.

[Radar WebKit Bug Importer]

<rdar://problem/173197678>

[Simon Fraser (smfr)]

Pull request: https://github.com/WebKit/WebKit/pull/62008

[EWS]

Committed 310575@main (0b5c61add5d5): <https://commits.webkit.org/310575@main>

Reviewed commits have been landed. Closing PR #62008 and removing active labels.

[EWS]

Committed 305413.615@safari-7624-branch (c1f3664d9ebb): <https://commits.webkit.org/305413.615@safari-7624-branch>

Reviewed commits have been landed. Closing PR #4896 and removing active labels.

[WebKit Commit Bot]

Re-opened since this is blocked by bug 311575

[Simon Fraser (smfr)]

It's not clear from specs about whether this is correct, so reverted.

[Simon Fraser (smfr)]

https://github.com/w3c/largest-contentful-paint/issues/152