From: Moore, Paul (paul.moore_at_[hidden])
Date: 2001-01-24 04:16:40
From: Darin Adler [mailto:darin_at_[hidden]]
> At some point I suggest we make the call for how we're going
> to do this throughout boost instead of constantly re-considering
> the issue separately for each part of the library.
I agree. My vote is very strongly in favour of assuming that injecting
overloads/specialisations/whatever into std:: is the correct approach, and
expecting user code to so this. This is based on pragmatism (it works well,
and it doesn't break any known compilers) as opposed to theory.
I think this effectively makes Koenig lookup irrelevant in the cases I care
about - which is good, as I can't rely on it being implemented...
This needs resolving, and the results documenting somewhere prominent in a
Boost policy document.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk