Re: [boost] [review] Review of Unordered library begins today December 7

From: David Abrahams (dave_at_[hidden])
Date: 2007-12-11 11:03:52

on Mon Dec 10 2007, "Daniel James" <daniel_james-AT-fmail.co.uk> wrote:

> On 10/12/2007, Thorsten Ottosen <thorsten.ottosen_at_[hidden]> wrote:
>> Daniel James skrev:
>> >
>> > This implementation actually does provide the no throw guarantee. But
>> > Jeremy Maitin-Shepard's version (on which it is based) didn't because
>> > the iterators didn't store the bucket, and required the hash to be
>> > calculated to work out which bucket the elements to be deleted are in.
>> > Which I think the standard allows.
>>
>> Ok, I think you right the standard allows that. But this is also
>> a bit irritating, because I cannot portable take advantage of your
>> stronger guarantee. feThis again begs the question of how much the
>> stronger guarantee affects performance?
>
> I don't think it affects performance at all, I think if it was
> required it would reduce the implementation choices. For example, if a
> container could rehash without invalidating iterators then I think it
> can currently do that when erase is called, but not if the stronger
> guarantee is required and hash can throw.
>
> Also, I made a mistake in the documentation, I said that a call to
> Hash or Pred can throw. But re-reading the standard, it looks like it
> doesn't have to be a call, it can be any member of the function
> objects.
>
>> yes, but my question was: *if an exception is thrown*, then what? I
>> assume you give the basic guarantee?
>
> Sorry, I misunderstood. This implementation gives the strong
> guarantee, the standard gives the basic guarantee.

I hope this implementation isn't sacrificing performance to get that
strong guarantee.

-- 
Dave Abrahams
Boost Consulting
http://www.boost-consulting.com