run_in_thread with no timeout uses 100% CPU

[hhuuggoo]

for pubsub, it seems redis-py supports non-blocking reads, blocking read, but not blocking with timeout? Would it be a good idea to add a timeout parameter that was passed through to the select call?

If so I can implement and submit a PR

[andymccurdy]

You can pass a float or int the sleep_time argument of run_in_thread. Every iteration of the loop calls time.sleep(sleep_time).

[hhuuggoo]

yea but that means you're guaranteeing a delay. If that's the best we can do than it's fine, but is there a way to do a blocking subscribe with timeout?

[andymccurdy]

Ah, I see what you're asking for now. That can probably be done by passing a timeout value to select.select as you suggest above.

[hhuuggoo]

ok thanks - I'll give this a shot an open a PR. If you have specific suggestions, please let me know, otherwise I'll give it a whirl

[github-actions]

This issue is marked stale. It will be closed in 30 days if it is not updated.