Re: [Boost-users] mpl question

From: Karl Meerbergen (Karl.Meerbergen_at_[hidden])
Date: 2005-09-23 09:04:21

On Friday 23 September 2005 15:35, David Abrahams wrote:
> Karl Meerbergen <Karl.Meerbergen_at_[hidden]> writes:
> > Hi Dave,
> >
> > The documentation at
> > http://www.boost.org/libs/mpl/doc/refmanual/vector.html
> > says for pop_front<v>::type:
> > A new vector of following elements: [next<
> > begin<v>::type>::type,end<v>::type)
> >
> > It suggests (in my opinion) that the result_type of pop_front is a
> > vector.
>
> It is a vector.  A vector does not necessarily have to be a
> specialization of mpl::vector.
>
> Hmm, I think in this case "vector" should not be set in code font,
> though.  Aleksey?
>
> > Should the documentation be clearer about this?
>
> That's what the use of "concept-identical" in the reference manual is
> intended to convey.
> http://www.boost.org/libs/mpl/doc/refmanual/extensible-sequence.html
>
> Do you have a specific suggestion that should be used to clarify the
> docs?

Perhaps it could be mentioned that pop_front, pop_back, insert etc return a 
VariadicSequence, (which it is I suppose?).
For me it would be clear it is not an mpl::vector type.
I am not enough familiar with the mpl concepts to be sure that my suggestion 
is correct.

>
> BTW, this point is also very clearly explained in "C++ Template
> Metaprogramming," in case you haven't read it.

I have read part of it. I agree that after a closer look, it is clearly 
explained.

Karl