Bug 305500 – REGRESSION(299380@main): [Skia] ref-test fast/text/text-combine-placement.html fails

• Component: New Bugs
• Status: RESOLVED FIXED

[Carlos Alberto Lopez Perez]

ref-test fast/text/text-combine-placement.html fails since 299380@main with this diff: https://build.webkit.org/results/WPE-Linux-64-bit-Release-Tests/305586@main%20(24590)/fast/text/text-combine-placement-diffs.html

Regressions: Unexpected image-only failures (1)
  fast/text/text-combine-placement.html [ ImageOnlyFailure ]


I checked to revert  299380@main locally and then the test passes.


The problem is that after 299380@main the tatechuyoko.svg is not loaded and it fallbacks to the system font.

In tatechuyoko.svg font the glyph for "A" and "の" are identical (are two squares)

$ grep -E 'unicode="(A|の)"' LayoutTests/fast/text/resources/tatechuyoko.svg
    <glyph unicode="A" horiz-adv-x="256" d="m 0 0 v 128 h 128 v -128 z"/>
    <glyph unicode="&#x306e;" horiz-adv-x="256" d="m 0 0 v 128 h 128 v -128 z"/>


So that is why the test passes when that font is loaded, but if isn't it fails because in the system fallback font the glyph for "&#x306e;" is the Japanese character "の"

[Carlos Alberto Lopez Perez]

This test was previously marked as "Failure" but that expectation is wrong for a ref-test.
A ref-test can't have a "Failure" result, because when it fails it will give instead a "ImageOnlyFailure" result (due to the pixel comparison between the test.html and the test-expected.html)
Expectations updated at 305610@main

[Carlos Alberto Lopez Perez]

Pull request: https://github.com/WebKit/WebKit/pull/56604

[EWS]

Committed 305664@main (546871d49b20): <https://commits.webkit.org/305664@main>

Reviewed commits have been landed. Closing PR #56604 and removing active labels.

[Radar WebKit Bug Importer]

<rdar://problem/168247259>