find_method changes and proposals

[Andrew Whitworth]

On Thu, Mar 4, 2010 at 11:32 AM, Allison Randal <allison at parrot.org> wrote:
> - 'find_method' looks up instance methods on Object, and class methods on
> Class/PMCProxy. That's intentional, a class in a metaobject protocol is
> itself a kind of object, and can be manipulated as an object.
>
> - The choice between op and vtable function boils down to a question of
> whether different PMCs need to be able to implement the behavior
> differently. For 'find_method' the behavior is different between Class and
> Object, alternate object models need to be able to implement it differently,
> and PIR subclasses need to be able to override it. So, it really needs to be
> a vtable function.

I think my last email response went only to chromatic, not to the list
at large. From the responses I have been getting from him and now this
response from you I think maybe I wasn't entirely clear with what I am
proposing.

I don't want to remove find_method or even really change it's
semantics in any way. find_method has the general semantic meaning
"find a method where the object is the invocant". What I want is to
also add a new VTABLE with semantic meaning "find a method where the
object is a metaobject for the invocant". I don't think such an
addition would break anything or even limit anything, but maybe I'm
not knowledgable enough on the topic to say.

I won't really argue the point any further, if it's not right it's not
right. Thanks for the input.

--Andrew Whitworth