Created attachment 271043 [details] a test case SSIA. Run the attached test case on threaded compositor enabled webkit build.
Created attachment 271150 [details] Patch
Comment on attachment 271150 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=271150&action=review Nice patch, thanks! Some nitpiks in the changelog.. > Source/WebKit2/ChangeLog:9 > + embeded environment. Probably main concern was increasing video memory for fixed elements. s/embeded/embedded/, I'm not sure whether is it correct reason or not. As we talked in IRC, it is because the fixed element is not considered to be a new stacking context at first. This feature was added later because of the scrolling performance in the embedded system. > Source/WebKit2/ChangeLog:10 > + However, it is time to rethink spliting the elements out to separate layers. Currently the fixed spliting/splitting and how about adding a comma after "Currently" ? > Source/WebKit2/ChangeLog:11 > + elements are redrawn to a backing store every time layout or scroll happens. It results in bad Also maybe you missed "the" before bad effect
Comment on attachment 271150 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=271150&action=review Thanks for your review! >> Source/WebKit2/ChangeLog:9 >> + embeded environment. Probably main concern was increasing video memory for fixed elements. > > s/embeded/embedded/, > I'm not sure whether is it correct reason or not. As we talked in IRC, it is because the fixed element is not considered to be a new stacking context at first. This feature was added later because of the scrolling performance in the embedded system. I see. Then Let's get rid of the whole lines. >> Source/WebKit2/ChangeLog:10 >> + However, it is time to rethink spliting the elements out to separate layers. Currently the fixed > > spliting/splitting > and how about adding a comma after "Currently" ? Yep. >> Source/WebKit2/ChangeLog:11 >> + elements are redrawn to a backing store every time layout or scroll happens. It results in bad > > Also maybe you missed "the" before bad effect Done.
Created attachment 271154 [details] Patch
Any concern?
I'm okay with this patch. Maybe cgarcia can help you. :)
Comment on attachment 271154 [details] Patch Thanks.
Comment on attachment 271154 [details] Patch Clearing flags on attachment: 271154 Committed r196623: <http://trac.webkit.org/changeset/196623>
All reviewed patches have been landed. Closing bug.