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

Subject: Re: [boost] [unordered] unordered_set::erase() complexity bug?
From: Joaquin M Lopez Munoz (joaquin_at_[hidden])
Date: 2010-03-17 17:29:39

Howard Hinnant <howard.hinnant <at> gmail.com> writes:

> 
> On Mar 16, 2010, at 2:27 PM, Brad Higgins wrote:
> 
> > 
> > On Mar 15, 2010, at 4:11 PM, Howard Hinnant wrote:
> >> I'm reviving an old thread with new information:
> >> 
> >> The LWG looked at this issue last week:
> >> 
> >> http://home.roadrunner.com/~hinnant/issue_review/lwg-active.html#579
> [...]
> Thanks for sharing your real-world experience Brad.  I've passed
> your note on to the C++ committee.  To answer your question, one
> vendor has already shipped these containers and with a design that
> reportedly does not suffer from this quadratic behavior, and is
> very reluctant to have this signature change.

Howard, I'm intrigued by the "reportedly" bit in the
sentence above. Has any member of the committee had access
to that code and assessed whether the quadratic behavior
is *effectively* avoided?

Thank you,

Joaquín M López Muñoz
Telefónica, Investigación y Desarrollo