Andrew Haley wrote: Mike Hore <mike_horeREM@ove.invalid.aapt.net.au> wrote: Mops/PowerMops used ... the GC stuff, just the basic vtable implemntation). Cheers, Mike. --------------------------------------------------------------- Mike Hore mike_horeREM@OVE.invalid.aapt.net.au --------------------------------------------------------------- ** Posted from http://www....
Mike Hore <mike_horeREM@ove.invalid.aapt.net.au> wrote: Mops/PowerMops used to only have early binding and Smalltalk-style late binding, but a few years ago it aquired references where you can put ref myClass myObj and then send messages to myObj. The compiler thus knows at compile time the base class of myObj (though it could be a subclass of myClass rather than ...
... wrote: John Passaniti <nntp@japanisshinto.com> wrote: Mike Hore wrote: So to get back to the idea of standardization,... just to explain my impending non-replies... :-) Cheers, Mike. --------------------------------------------------------------- Mike Hore mike_horeREM@OVE.invalid.aapt.net.au --------------------------------------------------------------- ** Posted ...
John Passaniti <nntp@japanisshinto.com> wrote: Mike Hore wrote: So to get back to the idea of standardization, I like the idea of low-level primitives on which anybody can build their own syntax. vtable and Smalltalk-style method dispatch are fundamentally different semantically, and as we've seen in this discussion there are plenty of valid uses for both. [...]...
Mike Hore has a good refcounting system in Mops as part of it's garbage collector. It makes it easier for his gc to figure out if an object is referenced from other objects or just itself. And the auto-release pools in Cocoa are a very good pure refcounting system, and you can, of course, combine the two approaches. Virtually, Michael Morris.
On Aug 7, 10:05 pm, Mike Hore <mike_hore...@OVE.invalid.aapt.net.au> wrote: I wrote: ..., I'll leave it there for now -- comments, anyone? Cheers,  Mike. ---------------------------------------------------------------    Mike Hore   mike_hore...@OVE.invalid.aapt.net.au --------------------------------------------------------------- ** Posted fromhttp://www.teranews....
Mike Hore wrote: So to get back to the idea of standardization, I like the idea of low-level primitives on which anybody can build their own syntax. vtable and Smalltalk-style method dispatch are fundamentally different semantically, and as we've seen in this discussion there are plenty of valid uses for both. [...] Are they really that different? My news feed has been ...
On Fri, 08 Aug 2008 12:35:47 +0930, Mike Hore <mike_horeREM@OVE.invalid.aapt.net.au> wrote: Nobody's replied yet (maybe everyone's kill-filed the OO syntax thread?) so I'm renaming the topic and going to do some rambling about possible standardizable primitives. We have four object systems to maintain, and clients who use them and will not abandon them because of a potential future ...
Bernd Paysan <bernd.paysan@gmx.de> wrote: Mike Hore wrote: So I think BIND actually needs to be three words, though I'm not sure what to call them. ?Let me, for now, call them EARLY-BIND VTABLE-BIND LATE-BIND Hm, given that there are already Neon implementations available that use table compactors and therefore would be ok with the vtable-bind primitive,...
