Subpixel layout: Remove float to LayoutUnit ceil/round function wrappers.

[zalan]

ceiledLayoutUnit()
roundedLayoutUnit()

These functions ceil/round float to LayoutUnit. Their usage are limited and can easily be confused with the snapping functions (snap to css/device pixels).

[zalan]

Created attachment 237156 [details]
Patch

[Simon Fraser (smfr)]

Comment on attachment 237156 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=237156&action=review

> Source/WebCore/platform/LayoutUnit.h:926
> +// When this occurs while accumulating converted LayoutUnit values, we can wind up with a result smaller than expected.

That doesn't explain the ceil though. This comment is pretty confusing. The caller really needs to know whether ceiling or flooring is appropriate for them.

> Source/WebCore/platform/LayoutUnit.h:927
> +inline LayoutUnit ceilFloatForSubPixelLayout(float value)

I don't think the "ForSubPixelLayout" adds anything. I think the name "ceiledLayoutUnit" is as good.

[zalan]

Created attachment 237158 [details]
Patch

[zalan]

Created attachment 237160 [details]
Patch

[Simon Fraser (smfr)]

Comment on attachment 237160 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=237160&action=review

> Source/WebCore/rendering/RenderBlock.cpp:2902
> +    preferredWidth = std::max(LayoutUnit::fromFloatCeil(result), preferredWidth);

You lost some of the comment here, which I think is useful to explain why this code uses ceil.

[WebKit Commit Bot]

Comment on attachment 237160 [details]
Patch

Clearing flags on attachment: 237160

Committed r172970: <http://trac.webkit.org/changeset/172970>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.