Bug 210732 – Content disappears on CSS parallax example

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Simon Fraser (smfr)]

Content disappears on CSS parallax example

[Simon Fraser (smfr)]

Created attachment 396942 [details]
Patch

[Simon Fraser (smfr)]

<rdar://problem/61997636>

[Simon Fraser (smfr)]

Created attachment 396943 [details]
Patch

[Darin Adler]

That new patch is a two-for-one!

[Simon Fraser (smfr)]

Oh whoops. I wish webkit-patch supported --git-index like prepare-Changelogs does.

[Simon Fraser (smfr)]

Created attachment 396945 [details]
Patch

[EWS]

Found 2 new test failures: compositing/tiling/coverage-adjustment-secondary-quad-mapping.html, fast/layoutformattingcontext/simple-absolute-positioned-replaced-inline-element-with-percentage-height.html

[Simon Fraser (smfr)]

https://trac.webkit.org/changeset/260371/webkit

[Darin Adler]

Comment on attachment 396942 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=396942&action=review

> Source/WebCore/platform/graphics/transforms/TransformState.cpp:223
> +    if (direction == ApplyTransformDirection) {
>          quad = m_accumulatedTransform->mapQuad(quad);
> +        return;
> +    }
>  
>      quad = m_accumulatedTransform->inverse().valueOr(TransformationMatrix()).projectQuad(quad, wasClamped);

I should have suggested else instead of return.

[Simon Fraser (smfr)]

I used return to mirror the mapPoint() code above.

[Darin Adler]

(In reply to Simon Fraser (smfr) from comment #10)
> I used return to mirror the mapPoint() code above.

I suggested else to mirror translateTransform above. We just happened to look at different functions.