On Mon, Sep 17, 2007 at 03:38:58PM -0700, Allison Randal wrote:
> PGE relies on an assumption of the old object metamodel that's no longer
> valid: that defining a method in the namespace of the class will allow
> you to call it as a class method.
> [...]
> $P0 = getclass 'PGE::Match'
> (mob, pos, target, mfrom, mpos) = $P0.'new'(mob, adverbs :flat :named)
>
> Under the new metamodel, this doesn't work, because instance methods and
> class methods are sharply distinguished.
> [...]
> I can think of several ways to work with this.
The one I'm likely to want to go with is to follow Perl 6's model,
which is that 'new' will be a method on a PGE::Match proto-object.
Indeed, I think that PGE also relies on Match objects themselves
being able to respond to 'new'.
So, the model I'd expect to ultimately aim for is something like:
$P0 = get_hll_global [ 'PGE' ], 'Match'
(mob, pos, target, mfrom, pos) = $P0.'new'(mob, adverbs :flat :named)
In this example it doesn't matter if $P0 is the actual class
object -- it can just be an instance of PGE::Match that
knows how to deal with the 'new' method (either because it's a
method in the class or because it's specifically composed into
the protoobject).
Most of the compiler tools are being created based upon this
model of object creation -- i.e., there's a proto-object
somewhere that is accessed via normal symbol table lookups,
and a 'new' method on that object is used to create an instance
of the class.
I can go ahead and switch PGE to use a model like this in the
trunk branch if that would help.
Note that PGE also uses a similar 'newfrom' method that has been
left in for compatibility with older versions of PGE -- eventually
those should also be converted into calls to 'new'.
Hope this helps,
Pm
