Bug 203642 – covscan: Incorrect check in RenderBox::hasOverrideContainingBlockContentHeight

• Component: WebCore Misc.
• Status: RESOLVED FIXED

[Eike Rathke]

Fix covscan identical branches

[Eike Rathke]

Created attachment 382388 [details]
Patch

[Alexey Proskuryakov]

Comment on attachment 382388 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=382388&action=review

> Source/WebCore/ChangeLog:11
> +        Checking HeightMap twice is definitely wrong.
> +        Given that hasOverrideContainingBlockContentHeight() should
> +        do the opposite of hasOverrideContainingBlockContentWidth() and
> +        correspond with overrideContainingBlockContentHeight(), do so.

Since this is a bug fix, it should come with a regression test that fails before the fix, and passes after the fix.

[EWS Watchlist]

Comment on attachment 382388 [details]
Patch

Attachment 382388 [details] did not pass win-ews (win):
Output: https://webkit-queues.webkit.org/results/13195733

New failing tests:
fast/repaint/backgroundSizeRepaint.html

[EWS Watchlist]

Created attachment 382465 [details]
Archive of layout-test-results from ews211 for win-future

The attached test failures were seen while running run-webkit-tests on the win-ews.
Bot: ews211  Port: win-future  Platform: CYGWIN_NT-10.0-17763-3.0.5-338.x86_64-x86_64-64bit

[Eike Rathke]

If the test now fails due to this change it means that cb->style().isHorizontalWritingMode() is false there. Really intended? i.e. does the tested code actually set that to false on purpose?

All other places that ask isHorizontalWritingMode() already use X/width for width if true, and Y/height if false (and vice versa for height), not doing it here looks unexpected to me. And then no check on isHorizontalWritingMode() at all would be needed.

The only place calling RenderBox::hasOverrideContainingBlockContentHeight() seems to be RenderBoxModelObject::relativePositionOffset()

I'm off here.

[Rob Buis]

This was fixed by https://bugs.webkit.org/show_bug.cgi?id=218504.