*) While we've facilities for a method cache, I'm ducking that for now. I expect performance to suck rocks until that's undone, but to do it right requires notifications.
*) While I *really* want to hide the details for now, under the hood we're going with a flat namespace system, with elements of names separated by NUL characters. If you're creating multilevel names (foo::bar::baz) then replace the separators with NULs. ASSEMBLY PROGRAMS ARE ALREADY ASSUMED TO DO THIS!
*) The name of the fallback method, which is called if no matching method is found, is referenced as the property "__FALLBACK__" on the class PMC. All languages generating parrot code should set this property appropriately. (Yes, I want to dodge the whole "what the heck is AUTOLOAD called in *this* language problem. Fooey. It's metadata, so we'll treat it as such)
*) We may or may not go with a split function/data namespace scheme. Probably not, and if we don't functions/methods will have a prefix character of some sort. Parrot will affix the appropriate prefix when it has to internally, so don't worry about this quite yet.
*) I've made the Parrot_base_vtable array movable again, as it needs to be resized. This is a temporary hack, since there are horrible threading issues here. (Not to mention the fact that this table is global and really ought not be, or we need a per-interpreter-family version that's actually used)
--
Dan
--------------------------------------"it's like this"-------------------
Dan Sugalski even samurai
[EMAIL PROTECTED] have teddy bears and even
teddy bears get drunk
