Bug 265957 – REGRESSION (271535@main): [ macOS Release wk2 ] fullscreen/full-screen-layer-dump.html is frequent failure

• Component: New Bugs
• Status: RESOLVED FIXED

[Marta Darbinyan]

Description:
fullscreen/full-screen-layer-dump.html is failing consistently on macOS Release wk2 started from 271535@main.

This issue can be bisected to 271535@main by using the command: 
run-webkit-tests  --release --iterations=50 fullscreen/full-screen-layer-dump.html

Confirmed:
271535@main - FAILED 50/50
271534@main - PASSED 50/50

History:
https://results.webkit.org/?suite=layout-tests&test=fullscreen%2Ffull-screen-layer-dump.html&version_name=Monterey&version_name=Sonoma&version_name=Ventura&flavor=wk2

Diff Log:
Test passes if you see a contents layer 3x the size and with a negative offset equal to the size:

(GraphicsLayer
  (anchor 0.00 0.00)
  (bounds 800.00 600.00)
  (children 1
    (GraphicsLayer
      (bounds 800.00 600.00)
      (contentsOpaque 1)
      (backgroundColor #000000)
      (children 2
        (GraphicsLayer
          (preserves3D 1)
          (children 1
            (GraphicsLayer
              (bounds 800.00 600.00)
              (children 2
                (GraphicsLayer
                  (anchor 0.00 0.00)
                  (bounds 800.00 600.00)
                  (contents layer (background color)
                    (position -800.00 -600.00)
                    (bounds 2400.00 1800.00)
                  )
                )
                (GraphicsLayer
                  (bounds 800.00 600.00)
                )
              )
            )
          )
        )
        (GraphicsLayer
          (preserves3D 1)
          (children 1
            (GraphicsLayer
              (bounds 800.00 600.00)


Link:
https://build.webkit.org/results/Apple-Sonoma-Release-WK2-Tests/271609@main%20(865)/fullscreen/full-screen-layer-dump-pretty-diff.html

[Radar WebKit Bug Importer]

<rdar://problem/119268470>

[EWS]

Test gardening commit 271641@main (b940f6dc0db7): <https://commits.webkit.org/271641@main>

Reviewed commits have been landed. Closing PR #21414 and removing active labels.

[Alexey Proskuryakov]

Looking at the history, it doesn't appear to be consistent, but is indeed failing very frequently. That must be why EWS didn't detect it.

[EWS]

Test gardening commit 271680@main (d57e747ddc89): <https://commits.webkit.org/271680@main>

Reviewed commits have been landed. Closing PR #21446 and removing active labels.

[Jer Noble]

It seems like this test is doing a layer dump without waiting for layout. There's a `setTimeout(..., 0)` but this will be insufficient to ensure that layout happens before dumping.

[Claudio Saavedra]

(In reply to Jer Noble from comment #5)
> It seems like this test is doing a layer dump without waiting for layout.
> There's a `setTimeout(..., 0)` but this will be insufficient to ensure that
> layout happens before dumping.

This test is now failing most of the times in most platforms (including Debug), causing false positive rejections in EWS. If a fix for the test is evident maybe it's worth fixing it? Otherwise it should get gardened not only for release.

[Claudio Saavedra]

Pull request: https://github.com/WebKit/WebKit/pull/21687

[EWS]

Test gardening commit 272043@main (de121adf537e): <https://commits.webkit.org/272043@main>

Reviewed commits have been landed. Closing PR #21804 and removing active labels.

[Marta Darbinyan]

Setting new test expectation since the test is failing on all configurations.

[EWS]

Committed 272053@main (ff8bb9f3e473): <https://commits.webkit.org/272053@main>

Reviewed commits have been landed. Closing PR #21808 and removing active labels.