From: joel de guzman (djowel_at_[hidden])
Date: 2002-04-10 00:51:05
----- Original Message -----
From: "Gennadiy Rozental" :
> --- "Andrei Alexandrescu" <andrewalex_at_[hidden]> wrote:
> >be avoided. To bring up the thing that led the first integration attempt to
> >failure, many people were convinced that loki should use mpl, while I
> >believe - then and now - that it would be better for loki to rely on a much
> >simpler typelist facility, and to export that typelist facility for use by
> >other boost libraries and by boost users.
> Given the fact that mpl is about to be review (I sincerely hope that you will participate), by the time when loki gonna be
reviewed typelist presumably will be part of boost already. So you will need really good argument for different implementation
(especially in boost public interface).
I tend to agree with Andrei. I would certainly hope to see a *smaller* typelist
library without all the bells and whistles. Such a library can be the least common
core subset that can be used by both mpl and loki, not to mention that
countless other libraries will benefit from it. I for one would love to use such
a library if it was small enough to fit in my pockets. The size and complexity
of mpl makes it unsuitable to use in the context of another library. 99% of
the time, we need just the basic stuff. Take for instance the named template
parameters. It would benefit a lot from such a small library yet it would be
foolish if such a facility will force me to drag the whole mpl along with it.
This of course assumes that mpl is a monolithic whole, for which I may be
wrong. It can be that mpl is just an orthogonal collection of modules and
we may use just the small subset without fear of unwanted interaction
with the rest of the library. So the question now is: how is mpl structured?
can we use a *tiny* subset of it without dragging in the whole library?
My 2c worth.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk