Re: [boost] File-Iteration Synopsis

From: David Abrahams (david.abrahams_at_[hidden])
Date: 2002-06-20 09:16:00

From: "Paul Mensonides" <pmenso57_at_[hidden]>

> Well, technically there isn't recursion is the horizontal mechanism
either. :)
>
> I used "iterate" precisely because it was inconsistent with
BOOST_PP_REPEAT (and
> clones).  In order to use a system like this effectively, the user really
has to
> know what is going to happen.

And how does "ITERATE" describe that any better than "REPEAT"?

> This typically requires a manipulation of the
> source order that makes it look more like the preprocessor is executing a
small
> program than preprocessing text.  Also, it isn't strictly necessary to
generate
> *text* with the mechanism.  I suppose you could make some other kind of
> calculation.

Not sure what you're getting at above.

> On this one, I like Aleksey's names better, which are the ones that I
used.

Can you explain why?

> I probably need to change the name of BOOST_PP_LOCAL_LIMITS though.

BOOST_PP_LOCAL_ITERATION_LIMITS (if you stick with ITERATION).

> Tip: for lists of small things, such as those
arg_tuple_size_helper-things from
> a while back, it is a preferrable to the use the simpler local iteration
system.
> Don' underestimate the usefulness of that facility even though it doesn't
have
> all the full-blown capabilities of the full-blown file-iteration.

I would always use the mechanism with the simpler interface unless there
was a good reason not to.