Dan Sugalski <[EMAIL PROTECTED]> wrote: > At 4:30 PM +0100 12/17/03, Leopold Toetsch wrote: >> getinterp P2 >> find_method P0, P2, "thread" >> find_global P6, "_foo" >> clone P5, P2 >> invoke # start the thread
> I think I'd like to start new threads a bit differently, at least > much more explicitly than an invoke. Invoke on an interpreter > arguably ought to dispatch to the interpreter *without* starting a > new thread. Its not invoking the interpreter, the "thread" method of a ParrotThread is invoked. Actually the example is a bit outdated - the clone is gone and a new .ParrotThread is instantiated. it should read now or finally: new P2, .ParrotThread set S0, "thread" # or better "start" or "run" find_global P5, "_foo" set P6, arg1 callmethcc It should be the same calling sequence as doing a method call on a new ParrotThread or any other object. But P2/P5 object passing did interfer. > We also need to work out the details of what and how we clone and > create new interpreters, what gets copied and what doesn't. I see > that thread's already started, though. :) Yep. Currently its the register file to get things running, nothing more. "Subject: threads and shared interpreter data structures" has an example of one of the problems I encountered til now. leo
