Re: [boost] combinations and permutations

Subject: Re: [boost] combinations and permutations
From: Howard Hinnant (howard.hinnant_at_[hidden])
Date: 2011-01-05 09:38:30

On Jan 5, 2011, at 4:41 AM, Mathias Gaunard wrote:

> On 04/01/2011 16:03, Howard Hinnant wrote:
> 
>>> I need to read this through thoroughly (it looks interesting), but along the same vein, an alternative abstraction is to create a range adaptor that allows iteration through permutations and combinations of a given range.  This would be more flexible than a for_each algorithm style of iteration, but possibly less efficient.
>> 
>> I'm not positive but I suspect that a range adaptor approach would suffer precisely the same inefficiencies I've demonstrated in the next_permutation/combination algorithms:  The increment operator begins to dominate as shown in the two figures.  That being said, the two approaches do not have to compete.  They can coexist.
> 
> You could generate the range adaptor from the for_each-style algorithm using a coroutine ; this should have minimal overhead.

I looked at:

http://en.wikipedia.org/wiki/Coroutine

and was unsuccessful in translating that idea to this application.  Perhaps you could elaborate?

-Howard