[webkit-reviews] review denied: [Bug 80550] Scrolling is slow on http://webkitmemes.tumblr.com : [Attachment 130716] Patch

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Mar 7 16:54:23 PST 2012


Simon Fraser (smfr) <simon.fraser at apple.com> has denied Anders Carlsson
<andersca at apple.com>'s request for review:
Bug 80550: Scrolling is slow on http://webkitmemes.tumblr.com
https://bugs.webkit.org/show_bug.cgi?id=80550

Attachment 130716: Patch
https://bugs.webkit.org/attachment.cgi?id=130716&action=review

------- Additional Comments from Simon Fraser (smfr) <simon.fraser at apple.com>
View in context: https://bugs.webkit.org/attachment.cgi?id=130716&action=review


> Source/WebCore/platform/graphics/Image.h:170
> +    // Supporting tiled drawing

I don't think the comment adds anything.

> Source/WebCore/rendering/RenderObject.cpp:1901
> +    updateSlowRepaintObjectState();

Seems a bit excessive to do this on every styleDidChange, and it might have
perf implications. Can we only do it when we know that style that affects the
slow repaint state changed?

If we did this then the bit might not be necessary.

> Source/WebCore/rendering/RenderObject.h:969
> +	   ADD_BOOLEAN_BITFIELD(isSlowRepaintObject, IsSlowRepaintObject);

Is it worth using a bit for this? What about keeping a hash? There won't be
many of these.

> Source/WebCore/rendering/style/FillLayer.cpp:132
> +static bool isImageSolidColorWithZeroAlpha(StyleImage* styleImage)

Hard to parse this name. Should it be isSolidColorImageWithZeroAlpha or
imageHasColorColorAndZeroAlpha?


More information about the webkit-reviews mailing list