[PATCH, v3] Bringing <mutex> and <condition_variable> up to spec.

[Benjamin Kosnik]

> Anyway, my basic points stand: what about the other threads
> implementations? Who is going to review that? ...

Does it make sense to add in dummy routines, as is done for
gthread-single, for the other thread models? Then expectations will
be consistent for all, and non-dummy versions can be added at a later
date if appropriate. 

That way the C++0x bits can move forward. 

May I suggest separating out the gthr* bits, and posting
that in a separate message? That part needs an OK from a global write
maintainer: as Paolo said, there is no thread maintainer (FWIW, Loren
used to do this AFAICT.)

The libstdc++ bits look ok to me: thanks for moving the ball forward on
this Chris.

Mark Mitchell seemed interested in reviewing these changes in the
immediate past: I would cc him and ask for comments.

best,
benjamin