"Andy Little" <andy@servocomm.freeserve.co.uk> writes:
"David Abrahams" wrote
"Arkadiy Vertleyb" wrote
result_of<minus_(int, _1_)>::type f = 3 - _1;
Does this direction make sense?
Any particular reason why BOOST_TYPEOF is not an option?
It requires type registration, whereas this approach should not. That's the only reason.
FWIW I would prefer to live with BOOST_TYPEOF and type registration. I have no doubt that registration will have its problems but only by familiarity will they be solved.
Huh? How will familiarity with the problem help? I can't ask my library's users to register types. Anyway, I can understand you saying "I'm content to live with...," but really, why would you _prefer_ a world where people like me, who aren't so content, can't get satisfaction?
Anyway somebody already proposed a similar looking alternative on comp.std.c++:
That proposal addresses something quite a bit different from mine. He's trying to get the result type of invoking a function, whereas I'm trying to get the type of the lazy function object that's created by Boost.Bind and/or Boost.Lambda. The difference is that the former can't be achieved without builtin typeof or registration, whereas the latter can. -- Dave Abrahams Boost Consulting www.boost-consulting.com