Re: Re: Inheritance graph

At 01:37 2003-08-14, David Abrahams wrote:
>Daniel Wallin <[email protected]> writes:
>
> > FYI, I've messed around with the code in inheritance.cpp /
> > inheritance.hpp and made it into a real singleton. In this process I
> > also investigated type importing, and to me it seems complicated to
> > try to share any data at all between modules.
>
>Yes, it's at least complicated.  However, given the size of the cache
>it does seem expensive not to share when multiple modules interact.
>
> > Also, I've looked into library configuration. I ended up with one macro;
> > BOOST_LANGBINDING_TYPE_INFO, which is the type used for type
> > identification.
>
>Is there some reason we'd have to use a MACRO instead of a typedef??

Well, the point was to allow for easier configuration. With a macro you can
generally just pass it to the compiler as a command line option, rather than
editing some library configuration header.

> > I also did something like this (python::type_info used for
> > illustrative purposes):
> >
> >    typedef BOOST_LANDBINDING_TYPE_INFO type_info;
> >
> >    template<class T>
> >    type_info static_type_id(T*, type<python::type_info>)
>                                 ^^^^^^^^^^^^^^^^^^^^^^^^^
>What is this argument for??

It's just a thought, configuration by overloading. Define the type_info type
with BOOST_LANGBINDING_TYPE_INFO and overload the type query
functions for that type.

I think we need to discuss the configuration solutions more.

---
Daniel Wallin



-------------------------------------------------------
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