Roy Stogner <> wrote on 08/26/2011 12:39:20 PM:

> On Fri, 26 Aug 2011, Kirk, Benjamin (JSC-EG311) wrote:
> > I'm happy to commit this - any objections?
> David Andrs has write permissions and can commit the patch himself.

Thanks for the review, both patches are in.

> If you got confused by the fact that he's asking for peer review
> before submitting an important patch, that's just because you've been
> dealing with jerks like me for too long.  ;-)
> And speaking of jerks and peer review:
> I haven't looked closely enough at patch 2, but at first glance
> everything looks good and if Ben trusts it then I surely will.
> For patch 1, we definitely need a fix there, since it turns out that
> this won't work with sets:
>    data.assign(vecdata.begin(), vecdata.end());
> But with some STL implementations it might be more efficient if we
> don't do a loop outside of their API.  Would it work to do this?
>    {
>      data.clear();
>      data.insert(vecdata.begin(), vecdata.end());
>    }

This worked just fine, so I changed the patch to do it this way.

> I think both sets and vectors support insertion of ranges.
> For that matter, there are two other places, lines 2371 and 2393,
> where I mistakenly assumed that std::set::assign existed.  We ought to
> give those methods the same treatment before someone gets around to
> instantiating them...

Fixed that as well.