From: Howard Hinnant (hinnant_at_[hidden])
Date: 2002-04-11 16:51:14
On Thursday, April 11, 2002, at 12:41 PM, David Abrahams wrote:
> 1. I'd like to have a few other people tell me I'm not crazy
I politely waited a long time. The silence is deafening and the results
are in. :-)
> 2. I think is_class can be reimplemented to avoid all of these
> problems using
> this technique:
> 3. Unfortunately, we'd need to keep the existing implementation of
> is_class for Pro7 support, since the sizeof() trick doesn't work there.
Fwiw, here's another possible implementation of is_class on Metrowerks:
Using this technique, even boost:is_union will work. But there may be
some minor differences between Metrowerks type_traits and boost
type_traits, such as facilities one lib has and the other doesn't.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk