Re: type traits, reflection, and whatnot

[Gabriel Dos Reis <gdr at integrable-solutions dot net>]

Phil Edwards <phil@jaj.com> writes:

| Testcases are mostly complete.  User documentation is complete, barring
| changes to the user-visiable names:
| 
|     http://www.devphil.com/~pme/reflection.html

Good!

I know that we will have to discuss any single extension we put there,
in particular in the way they diverge from the proposal voted on in
the TR, and how is the time (since we have no actual patches) but the
following comment confuses me:

  The GCC implementation of is_function will cause value to be true
  when T is a static member function, as well as the usual case of
  nonmemeber functions. Including static member functions may not have
  been the intent of the proposal.  


My understanding is that traits should be kept simple and straight.
static member functions are class-scope and not namespace-scoped
entities... so what preceisely is menat by function should be spelled
out clearly, not just mirroring an implementation detail.

-- Gaby