Re: [boost] [mpl]iter_fold_if Forward Backward rationale?

Subject: Re: [boost] [mpl]iter_fold_if Forward Backward rationale?
From: Larry Evans (cppljevans_at_[hidden])
Date: 2009-04-02 15:34:38

On 03/13/09 18:49, David Abrahams wrote:
 >
 > On Mar 9, 2009, at 10:00 AM, Larry Evans wrote:
 >
 >> Why is there both a ForwardOp and BackwardOp template argument to
 >> iter_fold_if?
 >>
 >> 
https://svn.boost.org/trac/boost/browser/trunk/boost/mpl/iter_fold_if.hpp#L62 

 >>
 >>
 >> I would have thought that only 1 was needed.
 >
 > Yeah, only one is needed if you are only interested in traversing the
 > list forward (or backward).
[snip]
 > Remember that list structures can only be built backwards, and
 > think about the use of inserters and such with the sequence
 > constructing algorithms.
 >

Lists can only be constructed backwards because lists don't have a
push_back?  IOW, with only a push_front, the lists have to be
constructed:

   push_front
   < push_front
     < push_front
       < list<>
       , i3
       >
     , i2
     >
   , i1
   >

to create list<i1,i2,i3>?  So, a traversal from back to front,
(i3->i1), is needed, and, for example,

   fold<vector<i1,i2,i3>,list<>, push_front<_,_> >

would give:

   list<i3,i2,i1>

Instead, to create list<i1,i2,i3> from some sequence, S<i1,i2,i3>,
reverse_fold would be needed:

   reverse_fold<S<i1,i2,i3>,list<>, push_front<_,_> >

Is that about what you mean?

[snip]
 >
 >
 > HTH,
 >

Yes. Thank you, David.

-regards,
Larry