Re: [boost] lifetime of ranges vs. iterators

From: Neil Groves (neil_at_[hidden])
Date: 2008-09-02 12:45:46

On Tue, Sep 2, 2008 at 5:32 PM, Arno Schödl <aschoedl_at_[hidden]> wrote:

> > optimum solution they are at least implementations of the functionality.
> I
> > would gladly accept superior contributions ;-). I am also happy to spend
> > more time improving them myself. I have to apologise for slow response
> > times, my full-time job simply doesn't allow me enough time to contribute
> to
> > this discussion as much as I probably ought to. You have made some very
> > valid technical points, let's focus on these please.
>
> My last post was not intended to put pressure on anyone to change anything,
> and I am very grateful for having boost. But Dave evidently thinks I am
> wasting everyone's time, and I beg to differ.
>

> If you agree that the general direction, self-contained iterators that use
> funky ranges to represent themselves efficiently, is right, I would bit by
> bit change my implementations of adaptor ranges and share them with you. I
> have not yet contributed to Boost, so probably my code is non-conformant all
> over the place, but it is a start.

I would really like to see some of the changes that you make, and that would
help me apply a consistent set of improvements to other ranges and adaptors
that could benefit. Once things are a bit more concrete we can measure and
compare the storage and performance benefits. There will be very little
argument with objective measurements. We can work together to make them work
and compliant with the intention of incorporating the changes into RangeEx.

>
>
> Arno
>
> --
> Dr. Arno Schoedl · aschoedl_at_[hidden]
> Technical Director
>

I hope this helps,
Neil Groves