Boost logo

Boost :

From: rogeeff (rogeeff_at_[hidden])
Date: 2002-02-23 17:47:42


--- In boost_at_y..., "Mark Rodgers" <mark.rodgers_at_c...> wrote:
> I've uploaded an initial version of auto_sequence to
>
> http://groups.yahoo.com/group/boost/files/auto_sequence.zip
>
> This follows Dave Abrahams's suggestion of turning my auto_vector
> into a container adapter.
>
> You can use it like this (oh for typedef templates):
>
> boost::auto_deque<foo>::type x;
> boost::auto_list<foo>::type y;
> boost::auto_vector<foo>::type z;
>
> Which are equivalent to
>
> boost::auto_sequence< std::deque<foo*> > x;
> boost::auto_sequence< std::list<foo*> > y;
> boost::auto_sequence< std::vector<foo*> > z;
>
> If you have some other sequence you wish to use, you could use
>
> boost::auto_sequence< my_nifty_sequence<foo*> > w;
>
> I've gone for the simple/unsafe approach of allowing you to modify
> the contained values:
>
> z[0] = new foo; // Whoops - memory leak.
>
> so Don't Do That. Personally I've never accidentally done this in
> many years of using similar "auto" containers so I don't think it
> is a problem. If anyone has strong opinions on whether there is
> a better alternative, please speak up.
>
> Please let me know if you think this, or something close to it,
> is worthy of documentation and submission.
>
> Thanks
> Mark

Why did you choose private inheritance? Why not public abd eliminate
all forwarding function?

Gennadiy.


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk