In https://bugs.webkit.org/show_bug.cgi?id=48282 I missed a few places where m_isGLES2 should be tested instead of calling gfx::GetGLImplementation().
Created attachment 73524 [details] Patch
Comment on attachment 73524 [details] Patch R=me
Committed r71793: <http://trac.webkit.org/changeset/71793>
There is a regression in this patch that causes worlds of webgl render incorrectly in in-process-webgl
Created attachment 73762 [details] patch: simple fix for the regression point-size.html should have detected this regression, only unfortunately due to mesa backend issues, this test is ignored in chromium bots.
Comment on attachment 73762 [details] patch: simple fix for the regression r=me
Committed r71942: <http://trac.webkit.org/changeset/71942>