|
Boost : |
From: John Maddock (john_at_[hidden])
Date: 2003-12-16 06:12:58
> > There was a major regression in this mornings tests. Looks like the
> > bool_trait_def.hpp change is the culprit.
> >
> > If someone (John?) fixes it, I'll rerun again later.
>
> This can be fixed by simply removing what John added.
> The correct thing to have added would have been:
>
> #include <boost/mpl/integral_c.hpp>
>
> template <class T, T val>
> struct integral_constant
> : mpl::integral_c<T,val>
> {};
>
> This would ensure compatibility with existing code that uses
> integral_c for tag dispatching.
Done for the primary template (currently unused), the problem was that
BOOST_TT_AUX_BOOL_TRAIT_SPEC2 didn't support non-type arguments on *some*
compilers, but did on those I used for testing (vc7.1). I'm still
rerunning the tests locally, but it should now be fixed.
John.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk