Re: Yet another multiple inheritence problem

Hi Eric,

thanks for concerning yourself with another obscure problem :)

> I was hoping to better understand what's going on with your patch
> before 
> responding, but that hasn't happened yet, so I figured I'd just
> respond.

Currently, the attached code just prints method resolution orders for
some class graphs (the second half of eieio-c3.el does that).

> I don't know anything about c3, but providing additional method 
> invocation orders sounds nifty to me.

If you agree, I would definitely want to incorporate it.

> An alternative possible solution if you don't get c3 "complete" as you
> like it 

I am currently working on changes concerning `eieiomt-method-list' and
`eieiomt-sym-optimize' to work with arbitrary class precedence orders.
However, despite being almost complete, the code is not currently in a
usable state. :(

> is that we could make eieio-default-superclass special in that
> it always shows up last on any invocation order.  In that way, you
> could 
> sort your own classes to avoid problems without worrying about the 
> default superclass.

That would be a nice improvement regardless of possible new method
resolution orders.

I hope to get the code for generic class precedence orders working
within a few days. Could we delay the decision until then?

Jan



------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev