[chromium] Clean up GraphicsContext3D initialization paths
Created attachment 128378 [details] Patch
Comment on attachment 128378 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=128378&action=review Looks good. r=me > Source/WebKit/chromium/ChangeLog:21 > + 2) GraphicsContext3DPrivate::createGraphicsContextFromWebContext() wraps the WebGraphicsContext3D in a Missing end of comment? > Source/WebKit/chromium/src/GraphicsContext3DPrivate.h:58 > + // via ContextLost before using the context in any way.. Once made current on a thread, the context cannot Two periods after "way".
Committed r108666: <http://trac.webkit.org/changeset/108666>
Committed r108679: <http://trac.webkit.org/changeset/108679>
Breaks tests. See commit above.
This broke some plugin tests because I wasn't properly retaining the temporary graphics context across WebViewImpl::graphicsContext3D() calls before the compositor was fully initialized. I've made an attempt at making this a bit clearer, because it sure confused the heck out of me.
Created attachment 128568 [details] delta from last patch
Created attachment 128571 [details] Patch
Ken - would you mind taking another look? The differences here are in WebViewImpl::graphicsContext3D() and WebViewImpl::createLayerTreeHost3D(). I've added some inline comments to try to explain what the code is trying to do a bit more explicitly.
Comment on attachment 128571 [details] Patch The delta looks fine to me.
Comment on attachment 128571 [details] Patch Clearing flags on attachment: 128571 Committed r108706: <http://trac.webkit.org/changeset/108706>
All reviewed patches have been landed. Closing bug.