Jenkins unhappy with high-precision elapsed time.

[sprinkmeier]

Jenkins parser appears to break when
test_case_attributes["time"] = "%f" % case.elapsed_sec
ends up with ludicrous number of decimal places.
Suggest using a format limit:
test_case_attributes["time"] = "%.3f" % case.elapsed_sec
(assuming milliseconds is considered enough precision)

[mlindner]

Some tests run very quickly, so the value should go to microseconds which is the minimum resolution of python datetime format anyway.

[sprinkmeier]

Jenkins breaks with anything more than millisecond precision.
I'm not sure more accuracy is useful.
If you're doing precision timing tests then you're probably better of running the test multiple times and averaging the result.

The formatting string could be made configurable so both use cases can be catered for.