Re: [Corelinux-develop] Thought, parser, template

[Frank V. C. <fr...@co...>]

Hans - Dulimarta wrote:
> 
> On Fri, 23 Feb 2001, Frank V. Castellucci wrote:
> 
> > Hans,
> >
> > As per previous conversation (corelinux, clfw, clfll), did you see
> > anything of interest? Is there something you want to do that isn't on
> > the list that makes sense for the direction of the libraries?
> >
> 
> My exploration in corelinux in not yet comprehensive. It is a new world to
> me, let alone the framework (clfw) and the library loader (clfll). All
> these three tracks are interesting to me, but I'll continue my exploration
> in the corelinux world first.
> 
> In the task list I found that #11180 and #11325 are unassigned.
> I'll take care of #11325 first, it has been past due for > 1 year :-)

Which one is 11325 again? I can't seem to get it on the web.

> The parser "project" looks cool too, I'm sure Christophe will enjoy
> working on it.
> 
> I have one thought about the metaclass for templated classes.
> I see the that DEFINE_METACLASS (...) might have to include the template
> type name. So, my idea is to add another DEFINE_... macro. But we also
> have to consider the number of template formal parameters. Perhaps, we
> could use several of these macros
> 
>    DEFINE_TEMPLATE_METACLASS1 (classname, tmpl1)
>    DEFINE_TEMPLATE_METACLASS2 (classname, tmpl1, tmpl2)
>    DEFINE_TEMPLATE_METACLASS3 (... , ...  , ...  , ...)
> 
> [probably we have to limit how many template parameters will be allowed.
> Refer to how the STL defines only AdaptableGenerator,
> AdaptableUnaryFunction, and AdaptableBinaryFunction]
> 
> Along with this, the Lisp-like meta declaration probably has to include
> some way of conveying the meta template parameter?
> 
> Examples:
> 
> template<typename T1>
> class A {
>   DEFINE_TEMPLATE_METACLASS1 (A, T1);
> ...
> };
> 
> template<typename X, typename Y>
> class B {
>   DEFINE_TEMPLATE_METACLASS (B, X, Y);
>   ...
> };
> 
> Will this work?
> 

Couple of things:

A templated declaration is useless until it has it's types defined, only
then is it a concrete class/object. The only way I have ever gotten
templated classes to work with generating their own meta-information was
to first create the concrete entity, and then the macro.

The whole idea of what Christophe is doing would be to do away with the
meta-macros, and instead use a more expressive and easy to read form for
which we generate the needed constructs. If you think about it, the
metatype/metaclass itself is mostly a collection of dictionaries
(relationships, data-members, methods, etc.).

The road into #define SOME_MACRO1() .... SOME_MACRO[n]() is exactley
what I avoided with the way I constructed the first round, NO LIMITS!!!
But, as you can see, it means more work on the developers part.

> --
> Hans Dulimarta, Ph.D. |                dul...@co...
> Research Associate    |     http://www.egr.msu.edu/~dulimart
> P: 517-432-7589       |     http://corelinux.sourceforge.net
> F: 760-281-7691     http://freshmeat.net/projects/snapsource
> Elec. & Comp. Engg., Mich. State Univ., E. Lansing, MI 48824
> 
> _______________________________________________
> Corelinux-develop mailing list
> Cor...@li...
> http://lists.sourceforge.net/lists/listinfo/corelinux-develop

-- 
Frank V. Castellucci
http://corelinux.sourceforge.net
OOA/OOD/C++ Standards and Guidelines for Linux
http://PythPat.sourceforge.net
Pythons Pattern Package