You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Searching for time.time() in the code base reveals many cases where we use it to compute durations, when it would be more robust to use a monotonic clock.
On Python 3.3+ this would be time.monotonic() or time.perf_counter() (or even time.process_time()), as appropriate.
Yeah, I've seen @beezz use monotonic clock in some ops code and it's good. This would maybe require a bit of care auditing all the time usage, as I'm not sure if we rely on the assumption on some deltas being >= 0, for example.
Searching for
time.time()
in the code base reveals many cases where we use it to compute durations, when it would be more robust to use a monotonic clock.On Python 3.3+ this would be
time.monotonic()
ortime.perf_counter()
(or eventime.process_time()
), as appropriate.On Python 2.x and <3.3 there is https://pypi.org/project/monotonic/.
See also https://www.python.org/dev/peps/pep-0418/.
The text was updated successfully, but these errors were encountered: