infact the expected behavior for the second example might also be: { t1 p1
t2 p2 t3 p3 } i am not sure about that part.
2011/4/29 Kenan Bölükbaşı <[email protected]>
> In this example, the generic word dispatches on p1 instead of t3:
>
> TUPLE: t1 s1 ;
> TUPLE: t2 < t1 ;
> TUPLE: t3 < t2 ;
> PREDICATE: p1 < t1 s1>> ;
> GENERIC: foo ( class -- print )
> M: t1 foo drop "t1" ;
> M: t2 foo drop "t2" ;
> M: t3 foo drop "t3" ;
> M: p1 foo drop "p1" ;
> t3 new t >>s1 foo
>
> ! --- Data stack:
> ! "p1"
>
> \ foo order
>
> ! --- Data stack:
> ! { t1 t2 t3 p1 }
> ! which probably should have been { t1 p1 t2 t3 } according to precedence
> rule and metaclass ranking rule combined.
>
> ! with a little scruffier structure
>
> TUPLE: t1 s1 ;
> TUPLE: t2 < t1 ;
> TUPLE: t3 < t2 ;
> PREDICATE: p1 < t1 s1>> ;
> PREDICATE: p2 < t2 s1>> ;
> PREDICATE: p3 < p2 s1>> ;
> t1 class-usages sort-classes
>
> ! --- Data stack:
> ! { t1 t2 t3 p2 p3 p1 }
> ! what i expected: { t1 p1 t2 p2 p3 t3 }
>
> ! with mixins
>
> TUPLE: t1 s1 ;
> TUPLE: t2 < t1 ;
> TUPLE: t3 < t2 ;
> MIXIN: m1
> INSTANCE: t2 m1
> MIXIN: m2
> INSTANCE: t3 m2
> t1 class-usages sort-classes
>
> ! --- Data stack:
> ! { t1 m2 m1 t2 t3 }
> ! i am not sure what was the expected behavior of this, i just wanted to
> test further. though i think there is a problem with this one, too.
>
> i am not experienced in OOP, but IMHO dispatch order on mixins should be
> handled specially, probably mixin should be placed in the order dynamically,
> i.e mixins placed in dispatch order every time according to which instance
> is of the mixin is making the dispatch predicate on the mixin true. I
> haven't tested for combinations with other metaclasses. but i can do some if
> needed.
>
> BTW, I wrote some words for tuple constructors aiming to handle shadowed
> slots better -with heavy help from folks in channel- :
> http://paste.factorcode.org/annotation.txt?id=1083 you might want to take
> a look at them. boa constructor normally tries to get values from stack for
> shadowed slots, too. shadowed-boa word first places the initial value of
> each shadowed slot at corresponding places in stack, (i didn't choose to
> just place f values because factor still checks for the declared slot type
> even when the slot is shadowed, so placing f doesn't work) then it calls boa
> as usual, and i wrote a mark-shadowed word that places a +shadow+ symbol as
> a value to those slots, mark shadowed uses low-level slot operators so it
> passes slot type checking. marking them makes it easy for user to see what
> is going on in the tuple if at inspector or listener. I am not sure if this
> would be something generally useful though, i hope it is. thanks.
>
------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network
management toolset available today. Delivers lowest initial
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd
_______________________________________________
Factor-talk mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/factor-talk
