[boost] Re: Re: Namespaces vs. directory placement

From: Jonathan Turkanis (technews_at_[hidden])
Date: 2004-12-29 13:36:04

Peter Dimov wrote:
> Jonathan Turkanis wrote:
>> The occurence of boost::io::put in the above example has to be
>> qualified, since otherwise it will refer to the member function being
>> defined.
>
> What I have to say has nothing to do with namespaces or directory
> placement, but...

> Given that in the documentation you speak of user-defined overloads of
> boost::io::read et al, it might be worthwhile to point out that
> two-phase lookup will not see these user overloads. If you want to
> have an overload customization point in a template, you have to use
> an unqualified call. And choose the identifiers wisely. :-)

Yikes!

I knew there were limitations to customization by overloading, but I didn't
think I had stumbled over them yet. I objviously can't use unqualified calls
here, or change the names.

I arleady implement read, et al. by simulated tag-dispatch:

    template<typename T>
    inline std::streamsize
    read(T& t, BOOST_IOSTREAMS_CHAR_TYPE(T)* s, std::streamsize n)
    {
        typedef typename detail::dispatch<
                    T, istream_tag, streambuf_tag, input
                >::type tag;
        return detail::read_impl<tag>::read(detail::unwrap(t), s, n);
    }

I'd really hate to make read_impl and dispatch public, or to add yet another
level of indirection. :(

So maybe my best option is to excise those innocent-sounding passages you refer
to, and require users to write wrappers if they want to integrate existing
components into the framework.

Thanks for pointing this out!

Jonathan