issues with Trellis event loops and threads

There's a problem with using the trellis and activity module in a
multithreaded app. There's a main loop thread with activity.EventLoop
running and there's a rule that runs periodically by rescheduling
itself (in my case it calls wx.ProgressBar.Pulse(), just to give you
an example of why this could be a legitimate use). There's also a
worker thread that initiates this process, but due to EventLoop and
Time being context.Service this doesn't work as expected -- the
dependency on _Timer instance (Time[0.1] for example) created from a
worker thread doesn't schedule a recalculation.

See test in attached file -- it should't hang up, but it does as
do_update rule only runs once.

I tried to fix it by restoring state from the main thread in the
worker thread but it's rather non-trivial as the State instances seem
to rely on a threadlocal data. If there's a proper way to do it, I'd
be happy to learn what it is. Thanks a lot in advance.

-- 
Best Regards,
Sergey Schetinin

http://s3bk.com/ -- S3 Backup
http://word-to-html.com/ -- Word to HTML Converter

_______________________________________________
PEAK mailing list
[email protected]
http://www.eby-sarna.com/mailman/listinfo/peak

from peak.events.trellis import *
from peak.events.activity import *
from threading import Thread

class ThreadedComponent(Component):
    updating = attr(False)
    iteration = attr(0)
    time_to_update = attr(EPOCH)

    @maintain
    def do_update(self):
        if self.updating and self.time_to_update:
            #print 'UPDATING'
            self.iteration += 1
            if self.iteration == 2:
                EventLoop.stop()
            self.time_to_update = Time[0.1]
            bool(self.time_to_update) # make sure we depend on new value



c = ThreadedComponent()

def start_updating_in_thread():
    c.updating = True

th = Thread(target=start_updating_in_thread)
th.start()
EventLoop.run()