[openlp-dev] New Thread API

[Raoul Snyman]

Hi folks,

Over the years we've used more and more threads in OpenLP for running 
things like web servers, downloads and other things that need to happen 
concurrently with the UI. Unfortunately because everything has been in 
bits and in various places, we've never had a unified way to run and 
keep track of threads. Thankfully we've largely not run into many 
issues, but more recently there have been segfaults not just when OpenLP 
stopped, but also preventing OpenLP from starting.

With this in mind I refactored how we run threads in OpenLP:

   from openlp.core.threading import ThreadWorker, run_thread

   class Worker(ThreadWorker):
       def start(self):
           """
           Do your stuff in here, then emit the quit signal.
           """
           self.server = Server()
           self.server.run()
           self.quit.emit()

       def stop(self):
           """
           If your thread is long-running, this is how OpenLP will stop 
it
           if it is still running when the user exits OpenLP.
           """
           self.server.stop()



   class MyClass(object):
       def run_server(self):
           """
           Run the server in a thread
           """
           worker = Worker()
           run_thread(worker, 'my_server')


I have also put together a wiki page explaining the threads API:

   https://wiki.openlp.org/Development:Threads


This change is not in trunk at the time of this e-mail, but the merge 
proposal is on Launchpad:

   
https://code.launchpad.net/~raoul-snyman/openlp/better-threading/+merge/335801


-- 
Raoul Snyman
+1 (520) 490-9743
raoul at snyman.info