From: Aleksey Gurtovoy (alexy_at_[hidden])
Date: 2000-04-30 18:22:09
Dave Abrahams <abrahams_at_[hidden]> wrote:
> Note that not having the 2-argument versions might prevent certain kinds
> generic code... not that I can really anticipate a concrete use. The point
> is that right now, you can write multipliable<T, U> and it will work even
> when T == U.
> That's really the only argument I can see that would carry much weight.
I thought you've already settled on the removal of the 2-type templates
versions without '2' at the end of names, and we are speaking only about
backward compatibility. In other words, I missed the fact that
implementation based on partial specialization has its own weight in context
of writing generic code. It's really may be the most important argument,
although at the moment I can't cite any concrete example either.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk