Migrating clients over to RImplGLES from GLESImpl would be easier if we could do it piecemeal (one subsystem at a time). This is conceptually simple because RImplGLES is a wrapper around a GLESImpl. However, one problem with that is that RImplGLES has a bit of internal state. This state would be invalidated if on part of the code uses GLESImpl in a way that RImplGLES is not tracking.
This bug is to see if we can remove the internal state from RImplGLES enough that it is safe to use concurrently with the GLES2Impl that it wraps.
Comment 1 by bugdroid1@chromium.org
, Oct 24