Re: [boost] [unordered] unordered_set::erase() complexity bug?

Subject: Re: [boost] [unordered] unordered_set::erase() complexity bug?
From: Stefan Strasser (strasser_at_[hidden])
Date: 2009-11-25 17:39:12

Am Wednesday 25 November 2009 21:54:54 schrieb Jeffrey Bosboom:
> Stefan Strasser wrote:
> > this is just an idea off the top of my head, I haven't thought this
> > through, but has the committee ever considered making functions
> > overloadable by their return type?
> >
> > void f(); // 1
> > int f(); // 2
> >
> > int main(){
> >   f(); //resolves to 1
> >   int b=f(); //resolves to 2
> > }
>
> That's a pretty drastic language change to solve an issue that is really
> a library problem.  
>

I don't want to argue for it because as I've said, I haven't really thought 
this through.
but this isn't a library specific problem, it applies to unnecessary 
evaluation of discarded return values in general.
it's just exacerbated in this case because you can't just add a parameter or 
create a second function like you normally would because the function 
signature is specified.

> (Besides, what if I want to call function 2 for its
> side effects and ignore its return type?  Even the ugly "(void)f()"
> doesn't seem to do what I want.)

I can't think of a reason why you'd want that, but 
(int) f(); 
should do the trick.