Re: [boost] lifetime of ranges vs. iterators

From: David Abrahams (dave_at_[hidden])
Date: 2008-08-29 09:56:18

on Fri Aug 29 2008, "Neil Groves" <neil-AT-grovescomputing.com> wrote:

>>
>> Aren't standard containers supposed to satisfy the Range concepts?  When
>> the container is destroyed, its iterators are invalidated, right?
>>
>
> Yes, containers satisfy Range concepts, but this question is more related to
> the opposite relationship. There is no need for a range destruction to
> invalidate the iterators, and this leads to simpler code for example when
> creating a temporary range to pass to a range algorithm.

So?  No algorithm could possibly _rely_ on the fact that the iterators
will be invalidated, so what does it matter?  You can do whatever's most
efficient.

-- 
Dave Abrahams
BoostPro Computing
http://www.boostpro.com