Re: [boost] metactrl header [compile time asserts]

From: David Abrahams (abrahams_at_[hidden])
Date: 2000-07-23 19:48:19

----- Original Message -----
From: "Greg Colvin" <greg_at_[hidden]>

> > BTW, I attended a Generic Programming conference with C&E in '98
> > http://www.cs.rpi.edu/~musser/gp/dagstuhl/gpdag.html where they talked
about
> > generative programming. Most of it seemed to be a pile of buzzwords and
> > excitement with little to back it up, but they did present a complete
lisp
> > interpreter implemented at compile-time with C++ templates!
>
> Now that's cool.  Got a link?  Or do I have to get the book?

No link, but I think I have the file somewhere (maybe). It will have to wait
until probably a few weeks after I move house at the end of the month,
though. That computer is packed.

> > From: "Greg Colvin" <greg_at_[hidden]>
> > > From: David Abrahams <abrahams_at_[hidden]>
> > > > I totally support all of Howard's arguments below :)
> > >
> > > Me too.  But what names to use?  Howard's names for 'IF'
> > > don't make it so clear what the meaning of the parameters
> > > is.  Maybe 'If' would do?
> > >
> > > Or if we are feeling very brave, how about something like
> > >    #define If whatever<
> > >    #define Then ,
> > >    #define Else ,
> > >    #define End >::type
> > > so that instead of the inscrutable
> > >    typedef If<flag, type1, type2>::type my_type;
> > > we can write
> > >    typedef
> > >       If flag Then
> > >          type1
> > >       Else
> > >          type2
> > >       End
> > >    my_type;
> > >
> > > And Howard thought he was getting punchy!
>
> No comment, eh?

Precisely!