|
Boost : |
From: Vladimir Ciobanu (psycho_at_[hidden])
Date: 2002-02-18 06:57:38
> Two usual suspects:
> - typedef templates
> - typeof (with gcc semantics?)
Yes, of course, these are usually on everybody's wishlist.
> What else should be on the list? For example, is there language support
that
> would assist type traits?
I've recently been in heavy discussions about this. The main problem, as
I see it is that the current C++ standard breaks one important rule of
programming: they seem to think about implementation in design stage. Of
course, thinking about implementation in terms of possible/impossible is
mandatory in design stage, but anything more is a mistake. Compilers can be
a lot smarter.
Somebody suggested reflection (RTTI+). That's one step ahead. But then
again, RTTI is not the answer to type traits at all. RTTI+ would presumably
be the answer for runtime type traits, if you agree calling it that.
However, I was also thinking about compile time reflection. A compiler can
easely detect things like is_base_and_derived, is_integral and other things
like this. This should be done via reflection.
Other than this, there are of course a lot of other problems which are
well known ( especially in the STL ) such as iterators or std::string.
PS: I'd like to see most of what Stroustrup had on his wish-list ( threads,
sockets, templated typedef's, the stupid little space we need when writing
templates "map this< is, weird< right> >", etc).
Vlad
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk