[Xcb] GSoC 2009: XKB and XI2 support

[Barton C Massey]

In message <20100120195800.GA7899 at feather> you wrote:
> > If we (as a project) are fine with protocol bits
> > existing before libxcb can do anything useful with them,
> > I can go ahead and push. But I have a sinking suspicion
> > that we might want to rev the protocol definition when
> > we get a better idea of what libxcb wants from it.
> 
> I agree with Peter's analysis here.  I really think we
> should not have protocol definition changes without
> implementation in at least one engine, ideally libxcb.

I don't know.  We used to do this all the time at the
beginning of the project, and it worked out fine.  When
someone came along and tried to use one of the protocol
definitions they found a few bad bugs, and they or we fixed
them.  What's the harm in having them in-tree?  We could
mark these descs as untested in a comment if your concern is
that people will use them unaware.

    Bart