When the root layer is changed, we should repaint the webview by calling syncLayers().
Created attachment 228399 [details] Patch
Comment on attachment 228399 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=228399&action=review > Source/WebKit/efl/ChangeLog:7 > + Maybe we need some description why? We discussed about regression. Is this patch for that? > Source/WebKit/efl/ChangeLog:11 > + (WebCore::AcceleratedCompositingContext::setRootGraphicsLayer): Deleted. This comment is ambiguous. > Source/WebKit/efl/WebCoreSupport/AcceleratedCompositingContextEfl.cpp:187 > + m_syncTimer.startOneShot(0); Don't we need to check wehther m_syncTimer is active?
(In reply to comment #2) > (From update of attachment 228399 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=228399&action=review > > > Source/WebKit/efl/ChangeLog:7 > > + > > Maybe we need some description why? > > We discussed about regression. Is this patch for that? Yes, I'll add some more description. > > Source/WebKit/efl/ChangeLog:11 > > + (WebCore::AcceleratedCompositingContext::setRootGraphicsLayer): Deleted. > > This comment is ambiguous. It's automatically generated comment. I'll replace it with "Move implementation from header file to source file". > > Source/WebKit/efl/WebCoreSupport/AcceleratedCompositingContextEfl.cpp:187 > > + m_syncTimer.startOneShot(0); > > Don't we need to check wehther m_syncTimer is active? We will set the syncTimer with 0 interval and do flushAndRenderLayers() as soon as possible even though syncTimer is active by css animations.
Comment on attachment 228399 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=228399&action=review >>> Source/WebKit/efl/WebCoreSupport/AcceleratedCompositingContextEfl.cpp:187 >>> + m_syncTimer.startOneShot(0); >> >> Don't we need to check wehther m_syncTimer is active? > > We will set the syncTimer with 0 interval and do flushAndRenderLayers() as soon as possible even though syncTimer is active by css animations. Well, if some javascript changes the condition of it to enable/disable sequentially? I am not sure whether it is impossible.
(In reply to comment #4) > (From update of attachment 228399 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=228399&action=review > > >>> Source/WebKit/efl/WebCoreSupport/AcceleratedCompositingContextEfl.cpp:187 > >>> + m_syncTimer.startOneShot(0); > >> > >> Don't we need to check wehther m_syncTimer is active? > > > > We will set the syncTimer with 0 interval and do flushAndRenderLayers() as soon as possible even though syncTimer is active by css animations. > > Well, if some javascript changes the condition of it to enable/disable sequentially? > I am not sure whether it is impossible. I understand your point. startOneShot(0) may be repeatedly called by setRootGraphicsLayer() calls. I'll add isActive() check.
Created attachment 228468 [details] Patch
Comment on attachment 228468 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=228468&action=review > Source/WebKit/efl/ChangeLog:10 > + Empty or non updated webview is shown when there is no update event after the root layer is changed. > + Because the syncTimer is not being triggered by root layer change. > + We should flush and render layers by starting syncTimer when the root layer is changed. Patch itself is fine to me. Nit, comment is bit ambiguous to me. (but I am also bad with english :) )
Comment on attachment 228468 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=228468&action=review > Source/WebKit/efl/ChangeLog:8 > + Empty or non updated webview is shown when there is no update event after the root layer is changed. I modify this description a slightly. "Even when there isn't any update event, empty(or, non-updated webview) webview can be shown after root layer is changed."
Created attachment 228476 [details] Patch
(In reply to comment #9) > Created an attachment (id=228476) [details] > Patch Good to me.
Comment on attachment 228476 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=228476&action=review r=me. Please land it after applying my comment. > Source/WebKit/efl/WebCoreSupport/AcceleratedCompositingContextEfl.cpp:185 > +{ Please add ASSERT(rootLayer); It looks rootLayer can be null when user set it as null by using ewk_view_root_graphics_layer_set().
Comment on attachment 228476 [details] Patch Clearing flags on attachment: 228476 Committed r166768: <http://trac.webkit.org/changeset/166768>
All reviewed patches have been landed. Closing bug.