Re: Relocation (= move+destroy)

[Marc Glisse <marc dot glisse at inria dot fr>]

On Thu, 18 Oct 2018, Marc Glisse wrote:

> On Thu, 18 Oct 2018, Jonathan Wakely wrote:
>
> > On 18/10/18 21:59 +0200, Marc Glisse wrote:
> > > On Mon, 15 Oct 2018, Jonathan Wakely wrote:
> > >
> > > > That's a good enough reason for me. Unless one of us implements
> > > > noexcept(auto) in GCC and Clang in the next few days, let's go with
> > > > what you have in the patch :-)
> > >
> > > Just checking: this 'ok' is for the noexcept style, it is not a review for 
> > > the complete patch yet, right?
> >
> > Right. I think I looked over the rest quickly and it seemed OK, but I
> > thought you were going to send a revised patch changing the triviality
> > test so was going to wait and review the revised patch. But since that
> > change is tiny I can still review the rest now anyway, so I'll do that
> > tomorrow.
>
> Ah, sorry for the misunderstanding, here is the version with is_trivial, 
> which regtested fine. Same ChangeLog.

Uh, why didn't I notice that the function __relocate is unused? I guess 
I'll resend the same patch without __relocate once retesting has finished 
:-( Sorry for all the corrections, I guess I didn't check my patch 
carefully enough before sending it the first time.

--
Marc Glisse