Subject: Re: [boost] [move] [range] move algorithm (was: interest: the pass-by-value...)
From: Krzysztof Czainski (1czajnik_at_[hidden])
Date: 2014-02-19 11:31:25
2014-02-19 12:10 GMT+01:00 Neil Groves <neil_at_[hidden]>:
> >> Why in the world does a stride iterator need to be random-access? Here
> >> is a strided range adaptor that works with input iterators:
> >> https://github.com/ericniebler/range-v3/blob/
> >> master/include/range/v3/view/stride.hpp
> >> It's not even hard.
> > I guess it may have been because it was the easiest. I see you have
> > special handling of bidirectional iterators. Anyway, nice to see your
> > improvements. :-)
> I think this implies that the current implementation in Boost.Range doesn't
> optimise for bidirectional and random-access which is wrong. My
> implementation requires at least a single-pass traversal but handles
> bidirectional and random-access fine.
Do we need a new Value Access Concept  for this: Readable Once Iterator?
The Iterator Traversal  is independent from Value Access.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk