Nested std::tr1::bind problem

[Paolo Carlini]

Peter Dimov wrote:
> The error that was posted indicated that variadic templates are being
> used; it seemed to me that if variadics are OK, so should be rvalue
> references, which is why I suggested the change.
Well, I would be tempted to say that, assuming the use of variadic
templates is user-visible, we should probably not have used variadic
templates in TR1. But this is after the fact, probably we didn't really
consider this issue, variadic templates are hugely useful for
compile-time, at the time didn't have in place a framework for
experimental C++0x facilities, etc. Right now, to be clear, I strongly
disagree with any non-trivial, non-regression fixing change to our TR1
implementation. If other maintainers want to approve work along these
lines, certainly I will not start an annoying litigation, I will simply
ignore that kind of work.

Paolo.