This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Simplifing tuples


chris wrote:

> Not sure if this is relevant, but I've been told that there is a good
> chance the next version of boost will convert their tuples from
> template metaprogramming to preprocessor magic.

Actually, what I knew is that the tuple was being rewritten using Fusion, a new
metaprogramming library of some sort. I'm CC'ing Joel, Fusion's (co?)author,
which could enlighten us.

Joel, if you want to catch up before replying, see the thread here:
http://gcc.gnu.org/ml/libstdc++/2004-10/msg00097.html. Basically, there were
worries that using too much template metaprogramming results in too complex
error messages (like what happens with the current version of boost::tuple<>).

As a way to reduce the impact of this, I am proposing a way to enhance the
concept check machinery with a new compiler extension, to be able to emit a
meaningful diagnostic within a static_assert-like construct. I would actually
like to hear comments from the v3 maintainers about this.

Giovanni Bajo



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]