Use render box background over border draw strategy in cases with background-image

[Justin Novosad]

Use render box background over border draw strategy in cases with background-image

[Justin Novosad]

Created attachment 176264 [details]
Patch

[Simon Fraser (smfr)]

Comment on attachment 176264 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=176264&action=review

> Source/WebCore/rendering/RenderBox.cpp:924
> +    // If there is only one layer and no image, check whether the background color is opaque
> +    if (!fillLayer->next() && !fillLayer->hasImage()) {
> +        Color bgColor = style()->visitedDependentColor(CSSPropertyBackgroundColor);
> +        if (bgColor.isValid() && bgColor.alpha() == 255)
> +            return true;
> +    }

What about an alpha image plus an opaque color?

[Justin Novosad]

(In reply to comment #2)

> 
> What about an alpha image plus an opaque color?

That could lead to color bleeding of the background color at the anti-aliased edges of rounded corners, so we want to use the TransparencyLayer strategy in that case.

[WebKit Review Bot]

Comment on attachment 176264 [details]
Patch

Attachment 176264 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/15025186

New failing tests:
fast/backgrounds/gradient-background-leakage-2.html

[WebKit Review Bot]

Comment on attachment 176264 [details]
Patch

Attachment 176264 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/15031059

New failing tests:
fast/backgrounds/gradient-background-leakage-2.html

[Justin Novosad]

Created attachment 177800 [details]
Patch

[Justin Novosad]

Review ping.

[WebKit Review Bot]

Comment on attachment 177800 [details]
Patch

Clearing flags on attachment: 177800

Committed r137473: <http://trac.webkit.org/changeset/137473>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Adam Klein]

Is there any chance that this would cause image diffs where background-image is not involved? I'm seeing slight diffs in the calendar picker after this change:

http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=platform%2Fchromium%2Ffast%2Fforms%2Fcalendar-picker%2Fcalendar-picker-appearance.html%2Cplatform%2Fchromium%2Ffast%2Fforms%2Fcalendar-picker%2Fcalendar-picker-appearance-ru.html%2Cplatform%2Fchromium%2Ffast%2Fforms%2Fcalendar-picker%2Fweek-picker-appearance-step.html%2Cplatform%2Fchromium%2Ffast%2Fforms%2Fcalendar-picker%2Fmonth-picker-appearance-step.html%2Cplatform%2Fchromium%2Ffast%2Fforms%2Fcalendar-picker%2Fweek-picker-appearance.html%2Cplatform%2Fchromium%2Ffast%2Fforms%2Fcalendar-picker%2Fmonth-picker-appearance.html%2Cplatform%2Fchromium%2Ffast%2Fforms%2Fcalendar-picker%2Fcalendar-picker-appearance-step.html

[Adam Klein]

Filed bug 104825 to track the image diffs in calendar-picker.