Re: [Adonthell-devel] Item spec

[Kai Sterker]

On Sun, 12 Jan 2003 16:55:16 +0100 Alexandre Courbot wrote:

> > With that I'll come to an end. Sorry if it's a lot of info, and not
> > all of it explained in detail. If nobody sees any huge flaws in the
> > above, I would go about implementing (and documenting) that stuff.
> > Sounds okay?
> 
> No time to read, as I'm in
> finals-panic-haven't-worked-enough-before-mode. But I know you enough
> to guess it must be ok in some way - and since we already discussed
> the matter before, I have my idea. :) How about adding this to the
> documentation (that is, using latex or doxygen)? Then you could rely
> on me to put some mess inside once I'm free! :)

Hm, I might do that before it gets lost on the mailing list again ;).
OTOH, I don't want to waste too much time and rather turn it into code
right away. Just wanted to give people like you and David a chance to
contribute :).


> Random idea: using Python attributes directly instead of a container
> of our own. That way an item directly "owns" its attributes, they can
> be of any type - and restricting the attributes or serializable
> objects will even make them totally saveable/loadable. This also
> implies an item is a Python object, nothing more, nothing less
> (expected of course a nice C++ interface).

That would be the right direction. With my current ideas I'm halfway
there, but still relying on py_object (and with basic attributes on C++
side for efficiency). Item-specific attributes are Python attributes
however, and each python item class has its put_state and get_state
method.

I'll have a look if it makes sense to add the python layer to the item
class directly without using py_object.

Kai