On Monday 28 June 2010 at 15:00, Jonathan Worthington wrote: > Certainly going in the namespace should not be the default though, which > is what iirc is what #389 fixed.
Exactly. This caused intractable problems for Rakudo. > Also the class registry probably should die. I don't know how to get rid of it. > A declarative PIR syntax for classes, attributes and methods is not what > Rakudo needs, though. It is, but you don't know it yet. Even if Rakudo doesn't *use* that syntax (and I recommend not using it), that syntax will force Parrot to come up with an efficient serialization and initialization mechanism which allows thawing of declared classes *without running initialization code to declare them*. > I fully expect Rakudo's class and role definitions to continue to > compile down to a series of method calls on meta-objects. However, what > should be possible is that we can then freeze the resulting object > (/meta-object) tree into the PBC, so that we can just bring it back to > life later on rather than having to run through re-creating it. Precisely. I have a page of notes on all of these needs, per the YAPC discussion. I want to run some design ideas past the Moose developers as well. -- c _______________________________________________ http://lists.parrot.org/mailman/listinfo/parrot-dev
