Subject: Re: [boost] [concept_check] Some issues
From: Robert Kawulak (robert.kawulak_at_[hidden])
Date: 2011-12-11 17:37:22
> From: Stewart, Robert
> Ah, here you've suppressed the compiler's ability to generate a default
> constructor, copy constructor, or copy assignment operator because the
> corresponding special functions are private in the base. Nice.
Actually, this is how it's already done in the current implementation too.
> > Is there anything wrong with combining archetypes?
> No, not if doing so does the right thing. Any implementation should
> strive to disable invalid combinations, even if invalid solely because
> of language or implementation limitations.
OK, but can you name such an invalid combination? I'm asking just out of
curiosity because I can't think of such a combination right now.
> > Sorry, but could you give a code example? I still don't see
> > your point.
> Obviously, I was arguing WRT the current implementation, not to one you
> had not yet shown and I didn't understand you to be proposing.
But my implementation doesn't differ from the current one in any significant
way other than I've already explained (i.e., it removes the superfluous
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk