Remove algo duplications

[Marc Glisse]

On Tue, 19 Jun 2012, François Dumont wrote:

>    No opinion regarding this patch ? Here is a slightly modified version 
> without usage of lambda in is_heap_until implementation so that we use move 
> semantic when possible.

Hi,

(not a comment on your patch) did you have any opinion about the 
possibility to use indirect functors? I.e. functors that get called as 
f(i,j) instead of f(*i,*j), and f(a,b) typically returns *a<*b. I am not 
sure it is a good idea, but it might avoid any regression.

-- 
Marc Glisse