Bug 101974 – Boxes with rounded corners and thin borders are too slow to draw

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

[Justin Novosad]

Boxes with rounded corners and thin borders are too slow to draw

[Simon Fraser (smfr)]

Testcase?

[Justin Novosad]

Created attachment 173887 [details]
Patch

[Justin Novosad]

The expected pixel results for the following layout tests will require rebaselining due to minor discrepancies along anti-aliased rounded corner edges:
fast/borders/border-radius-wide-border-01.html
fast/borders/borderRadiusInset01.html
fast/forms/validation-message-appearance.html
svg/custom/svg-fonts-in-html.html

[WebKit Review Bot]

Comment on attachment 173887 [details]
Patch

Attachment 173887 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/14815638

New failing tests:
fast/borders/border-radius-wide-border-05.html
fast/forms/validation-message-appearance.html
svg/custom/svg-fonts-in-html.html
fast/borders/border-radius-wide-border-01.html
fast/borders/borderRadiusInset01.html

[Simon Fraser (smfr)]

Comment on attachment 173887 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=173887&action=review

> Source/WebCore/ChangeLog:9
> +        With the current implementation RenderBox::
> +        determineBackgroundBleedAvoidance() the slow path

I think you're missing a 'uses'

> Source/WebCore/ChangeLog:12
> +        borders.  This is because the BackgroundBleedShrinkBackground

One space after periods.

> Source/WebCore/ChangeLog:18
> +        By using this approach rater than BackgroundBleedUseTransparencyLayer,

rather

> Source/WebCore/rendering/RenderBox.cpp:897
> +bool RenderBox::backgroundIsSingleOpaqueLayer() const

Multiple sold color backgrounds would also be OK, no?

> Source/WebCore/rendering/RenderBoxModelObject.cpp:703
> +    adjustedRect.inflateY(-static_cast<LayoutUnit>(ceil(1 / transform.yScale())));    

Whitespace at EOL?

> Source/WebCore/rendering/RenderBoxModelObject.cpp:713
> +        adjustedRect.inflateX(-static_cast<LayoutUnit>(ceil(1 / transform.xScale())));

We should probably provide a helper function for this shrinking.

> LayoutTests/ChangeLog:12
> +        * fast/borders/border-radius-wide-border-05.html: Added.

Where's the pixel result?

> LayoutTests/fast/borders/border-radius-wide-border-05.html:11
> +        -moz-box-sizing: border-box;

Remove.

[Build Bot]

Comment on attachment 173887 [details]
Patch

Attachment 173887 [details] did not pass mac-ews (mac):
Output: http://queues.webkit.org/results/14809933

New failing tests:
fast/borders/border-radius-wide-border-05.html

[Justin Novosad]

(In reply to comment #5)
> Multiple sold color backgrounds would also be OK, no?
> 

Almost.  To do it right we'd need logic to optimize-out the bottom layers.  Otherwise, if we just draw the multiple solid layers on top of each other, there will be color bleeding along anti-aliased edges.

[Justin Novosad]

Created attachment 173968 [details]
Patch for landing

[WebKit Review Bot]

Comment on attachment 173968 [details]
Patch for landing

Rejecting attachment 173968 [details] from commit-queue.

Failed to run "['/mnt/git/webkit-commit-queue/Tools/Scripts/webkit-patch', '--status-host=queues.webkit.org', '-..." exit_code: 2

Last 500 characters of output:
ucceeded at 1 with fuzz 3.
patching file LayoutTests/fast/borders/border-radius-wide-border-05-expected.txt
patching file LayoutTests/fast/borders/border-radius-wide-border-05.html
patching file LayoutTests/platform/chromium/TestExpectations
Hunk #1 FAILED at 3599.
1 out of 1 hunk FAILED -- saving rejects to file LayoutTests/platform/chromium/TestExpectations.rej

Failed to run "[u'/mnt/git/webkit-commit-queue/Tools/Scripts/svn-apply', u'--force']" exit_code: 1 cwd: /mnt/git/webkit-commit-queue

Full output: http://queues.webkit.org/results/14838054

[Justin Novosad]

Created attachment 174151 [details]
Patch for landing

[Justin Novosad]

Comment on attachment 174151 [details]
Patch for landing

Forgot to "svn resolved TestExpectations".  Must resubmit.

[Justin Novosad]

Created attachment 174189 [details]
Patch for landing

[WebKit Review Bot]

Comment on attachment 174189 [details]
Patch for landing

Rejecting attachment 174189 [details] from commit-queue.

Failed to run "['/mnt/git/webkit-commit-queue/Tools/Scripts/webkit-patch', '--status-host=queues.webkit.org', '-..." exit_code: 2

Last 500 characters of output:
um/third_party/angle --revision 1389 --non-interactive --force --accept theirs-conflict --ignore-externals' in '/mnt/git/webkit-commit-queue/Source/WebKit/chromium'
34>At revision 1389.

________ running '/usr/bin/python tools/clang/scripts/update.py --mac-only' in '/mnt/git/webkit-commit-queue/Source/WebKit/chromium'

________ running '/usr/bin/python gyp_webkit' in '/mnt/git/webkit-commit-queue/Source/WebKit/chromium'
Updating webkit projects from gyp files...
Total errors found: 0 in 1 files

Full output: http://queues.webkit.org/results/14824840

[Justin Novosad]

Created attachment 174201 [details]
Patch for landing

[WebKit Review Bot]

Comment on attachment 174201 [details]
Patch for landing

Clearing flags on attachment: 174201

Committed r134631: <http://trac.webkit.org/changeset/134631>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.