Re: Registries [2]

Daniel Wallin <[email protected]> writes:

> At 15:12 2003-07-31, David Abrahams wrote:
>>Daniel Wallin <[email protected]> writes:
>>
>> > At 03:08 2003-07-31, David Abrahams wrote:
>> >>Daniel Wallin <[email protected]> writes:
>> >>
>> >> > At 00:25 2003-07-31, David Abrahams wrote:
>> >> >
>> >> >>This seems to me to be the killer argument against having any implicit
>> >> >>imports of from-xxx converters.  No?
>> >> >
>> >> > Well, yeah.. Unfortunately the same argument applies to the
>> >> > cast-graph too.
>> >>
>> >>:(
>> >
>> > But we can still share the code though, right?
>>
>>Sure.
>>
>> > I already made the local registry into a singleton, so we'd just do
>> > the same for the inheritance-code and link registry<T>::instance()
>> > and inheritance_graph::instance() statically.
>>
>>And how would two extensions share a cast-graph?
>
> They wouldn't. 

That would be bad.  You need to be able to write derived classes in
different extensions from the base.

> Perhaps it would be possible to share subgraph's without duplicating
> the data, but I don't know.
>
> Or did you mean how it would be related to importing? In that case I
> guess we'd just import all reachable vertices from the vertex
> associated with the imported type..

Hmm, you're suggesting importing cast graph components?  I don't know;
it seems like a waste for several extensions which are known to
collaborate to have to each maintain a separate copy of the cast
graph.  I think we can do better.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01