On Thu, Jan 27, 2011 at 12:03, Daniel da Silva <ddasilva@umd.edu> wrote:
We have a threading.Timer class which executes after a given delay and then terminates. I was thinking about a pre-made class that would execute, wait an interval, and then repeat. It would cut down on the logic people need to implement themselves, and make simple scripts faster.

Here are some use cases:

# Send a ping message to all connected clients every 120 seconds
from server import ping_all_clients
pinger = threading.RepeatTimer(120, ping_all_clients)

# Check for updates every 3 hours
from mymodule import check_for_updates
update_checker = threading.RepeatTimer(60*60*3, check_for_updates)

I was thinking of the class having an initializer signature as follows:

class threading.RepeatTimer(interval, function, args=[], kwargs={}, limit=None)
     Create a timer that will run function with args and kwargs and repeat every interval secods. If limit is
     an integer, limits repetitions to that many calls.

         Stop the repeat timer. If in the middle of a call, the call is allowed to finish.


I'm not sure this is a good fit in the standard library. There's too much that the implementer might want to customize for their repeating needs. For your pinger, what if a ping fails? Keep pinging while it fails? Timeout for a bit and try again? Increasing timeouts? Just exit? I don't think there's a general enough use case that a repetitive timer has in order to be included.

The following will likely do what you want.

import threading

class RepeatTimer(threading.Thread):
    def __init__(self, interval, callable, *args, **kwargs):
        self.interval = interval
        self.callable = callable
        self.args = args
        self.kwargs = kwargs
        self.event = threading.Event()

    def run(self):
        while self.event.is_set():
            t = threading.Timer(self.interval, self.callable,
                                self.args, self.kwargs)

    def cancel(self):