Code review comment for lp://qastaging/~mir-team/mir/animated-cursors

> Its rather, the existing BufferStream has special bits about being a surface,
> aka SurfaceBufferStream.

That sounds like SurfaceBufferStream /IsA/ BufferStream and the abstractions need fixing.

> In the global picture though it makes more sense to
> me to let that remain as "BufferStream" though.
>
> I think maybe the biggest misnomer is mir::scene::SurfacelessBufferStream, but
> the BufferStream doesn't appear in the scene! Perhaps it should be
> mf::BufferStream(Impl?)?

It would still make more sense to me that the scene contains a variety of objects some of which have associated buffer streams. But I don't think the scene needs to know anything more than "there is a type BufferStream". Having surface derive from BufferStream makes that impossible.

Of course, that may well require more change to existing code than what is proposed. I've not investigated in enough depth.