Re: Our 4.0 Changes

[Paolo Carlini <pcarlini at suse dot de>]

Chris Jefferson wrote:

> Feel free to not put these things in, they are just the things that
> occur to me, of things I have been involved in :)
>
> std::iter_swap (and therefore most of the mutating algorithms) now
> makes an unqualified call to swap when the value_type of the two
> iterators is the same. While not required by the standard, iter_swap
> will still work on proxy iterators as long as their their
> reference_type is not "value_type&".

Agreed, but maybe something more concise is more suited. Will look into
that.

> Also, do you want to add something like:
>
> No guarantes about the implementation of the code in TR1 is provided.
> In particular the final location of the headers and namespace in which
> they will be in has not been finalised, and it is not promised that
> libstdc++-v3 will remain link compatable when code using TR1 is used.
>
> The link-compatable bit in particular I think about.. in particular
> tuple<> won't be link compatable if next version we give it a maximum
> of 20 or 30 parameters, and if anything changes that might break
> things to...

Yes, this is absolutely right, but (see above) we should aim for
something a little shorter, I think.

Thanks!
Paolo.