Re: [Patch] libstdc++/32907

[Benjamin Kosnik <benjamin dot kosnik at gmail dot com>]

> the below is the minimal patch which I prepared and tested for this 
> performance PR. Per the standard, in general the equality operator is 
> required to simply return string::compare. However, as always, for 
> non-user defined types we have leeway, because the user cannot
> provide is own specialization anyway (we are already exploiting that
> in many other places, of course). As regards the implementation
> details, Iknow we don't have many enable_ifs in user-visible code (we
> have some in cmath, for example, but here I can definitely do
> without, if people care), but I think this is an appropriate case,
> which also scales well if we want to extend the implementation to
> additional character types on the path to C++0x.

Hey Paolo. Looking good.

I would like to see a serious effort on behalf of the new character
types, as soon as possible. 

If you do the stuff from Matt's Oxford paper, I'll do the code
conversion stuff from PJP's paper in Toronto.

-benjamin