[Ecls-list] ECL + ContextL?

[Juan Jose Garcia-Ripoll]

On Mon, Nov 2, 2009 at 8:11 AM, Alexander Gavrilov <angavrilov at gmail.com> wrote:
> Here is an online version of the spec (or so it seems):
>
>  http://www.lisp.org/mop/concepts.html#inheritance
>
> This table specifies both that 'class' is an abstract
> class, so there should be no instances of it, and that
> funcallable-standard-class inherits directly from class.

Too bad. Then we might have to fake a common ancestor.

> On the other hand, funcallable-standard-object does seem
> to be a subclass of standard-object.

I must admit I sometimes find the AMOP mind-bogling.

> Also, a comment in kernel.lsp says that standard-generic-function
> must be 'first in the class hierarchy'? Does that mean that it must
> actually inherit directly from standard-object, or only that all
> ancestors must not have any slots of their own? The spec says
> that it should inherit from a bunch of abstract classes, e.g.
> funcallable-standard-object.

No, it simply means that if you use (defclass foo (class1 class2 ...
classn) ...) and generic-function is in that list of ancestors, then
it must be class1 The reason is that we need some control on the slot
positioning for ECL's dispatch function to work.

Juanjo

-- 
Instituto de Física Fundamental, CSIC
c/ Serrano, 113b, Madrid 28006 (Spain)
http://juanjose.garciaripoll.googlepages.com