Subject: Re: [boost] [multi_index] Move Semantics
From: Fernando Pelliccioni (fpelliccioni_at_[hidden])
Date: 2013-02-27 05:00:27
On Wed, Feb 27, 2013 at 4:22 AM, Joaquín Mª López Muñoz <joaquin_at_[hidden]>wrote:
> El 26/02/2013 14:09, Fernando Pelliccioni escribió:
>> I'm using quite Boost.MultiIndex and I would like it has support for
>> I am willing to devote time to it if nobody is doing it.
>> Joaquin has recently stated that hi has changes to commit to one of his
>> libs. ( Thread: "Should I commit to SVN or Git?" )
>> I wonder if those changes are related to add C++11 support to
> Hi Fernando and team,
> The latest changes I commited to the lib are unfortunately not related
> to C++11 upgrading but very minor patches to suppress warnings and such.
> Boost.MultiIndex sorely needs to have move semantics, initializer list
> construction and emplace functions. Even though I'm willing (and have
> been for some months) to add these myself the sad reality is that the
> time I can devote to this is increasingly small, so if any other
> contributor wants to take up the task I'll be more than happy to assist.
> A few words on the changes:
> * In general, when implementing a new feature I've had more work
> upgrading tests and docs that writing the feature itself. Many patch
> contributors forget to provide the former, though.
> * Implementing move construction and initializer list support should be
> straightforward. Probably the design decisions in Boost.Unordered (and
> part of the code itself) can be reused here.
> * emplace is a different animal: apart from the interface implementation
> (which, again, might benefit of the work already done in
> Boost.Unordered) some internal scaffolding needs to be put in place.
> I'll be happy to discuss this.
> Joaquín M López Muñoz
> Telefónica Digital
I will try to send both patches and updates to the test. Regarding to the
documentation perhaps we should work together.
Casually, one of the first changes I thought implement is emplace, I was
not aware of the problem you mention.
I will check the design decisions taken in Unordered and Container, in any
case we can discuss it.
Thank you very much for your advice and keep in touch.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk