time.sleep(1) sometimes runs for 200 seconds under windows

[Paul Rubin]

Dennis Lee Bieber <wlfraed at ix.netcom.com> writes:
> 	Any chance they all tend to have slow clocks, and are getting bitten
> by a semi-random NTP time update; my machine tends to run the NTP update
> at 7-day intervals (including time of day), counting from the last
> successful synchronization.

This is a good possibility, that the clock is jumping ahead rather
than that 200 seconds of actual real time is going during the
time.sleep call.

What happens if you replace the time.time call with something that
reads the current time from an SNTP server (or equivalent) through a
socket?