Optimize outline rendering to avoid transparency layers

• Component: Layout and Rendering
• Status: RESOLVED FIXED

[Ben Wells]

Simple cases of borders and outlines, such as where all sides are the same color, are SOLID style, and don't have rounded corners, could easily be optimized to avoid using expensive transparency layers.

[David Barr]

Created attachment 113617 [details]
Patch

[David Barr]

Created attachment 113618 [details]
Patch

[WebKit Review Bot]

Comment on attachment 113618 [details]
Patch

Attachment 113618 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/10314046

New failing tests:
fast/box-shadow/box-shadow-radius.html
fast/layers/self-painting-outline.html

[David Barr]

> New failing tests:
> fast/box-shadow/box-shadow-radius.html
> fast/layers/self-painting-outline.html

These will likely need to be rebaselined due to Skia rounding differently for layered vs direct alpha paints.

[Darin Adler]

Comment on attachment 113618 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=113618&action=review

> Source/WebCore/rendering/RenderObject.cpp:1145
> +    if (outlineStyle == SOLID && outlineColor.hasAlpha()) {

I think this code needs a why comment.

If you look at just this code, it's surprising that we have a special case just for the case where we have alpha. A good way to make that clear is to put this optimization inside the "use transparency layer" if statement. There it is easy to see that we are avoiding an expensive transparency layer. Otherwise, it is unclear why there is no fast case for the case where we don’t have alpha. The reason is that the code is already fast, and this alternate implemenation would probably be slower, in cases where we are not making a transparency layer.

So you could just add a comment, or you could even put this entire fast case inside the useTransparencyLayer if statement, or you could at least use the useTransparencyLayer boolean to trigger this fast case.

I think the key is that this is an optimization because it avoids the transparency layer and the code should make this crystal clear.

[David Barr]

Created attachment 113752 [details]
Patch

[David Barr]

Comment on attachment 113752 [details]
Patch

Moved the short path inside the useTransparencyLayer test for clarity.
Updated pixel tests for chromium-linux due to subtle change in alpha rounding with direct paint vs transparency layers.

[WebKit Review Bot]

Comment on attachment 113752 [details]
Patch

Clearing flags on attachment: 113752

Committed r99454: <http://trac.webkit.org/changeset/99454>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[noel gordon]

Rebaselines added in http://trac.webkit.org/changeset/99529