freezegun

This is a hard-earned lesson, but it's true. Instead of doing def do_something_with_date(): now = datetime.now() return now - timedelta(days=2) you should do def do_something_with_date(now): return now - timedelta(days=2) and explicitly pass in edge-case dates into the `now` param in your unit tests. Alternatively, if you're using Python, use the freezegun library to fix the current time in tests: https://github.com/spulec/freezegun

* in C++. On Python, just use freezegun to inject controllable timestamps in response to calls to time methods. https://github.com/spulec/freezegun

So I guess your units tests run with the current date and don't test behavior with different dates. That's something which has the potential to hide quite a few date related bugs until it's too late. What I like to do to avoid this situation is to use property-based testing to test that the code works for all dates in the desired time span. For Python I use hypothesis [1] in combination with freezegun [2] for that. Here is an example of how that could look like: from datetime import datetime, timezone as tz from freezegun import freeze_time from hypothesis import given, strategies as st def business_logic(): # TODO: Refactor before 2038! if datetime.now(tz=tz.utc) > datetime(2038, 1, 19, 3, 14, 7, tzinfo=tz.utc): raise Exception("This code is not working after UNIX time overflow") ...

I have two main use cases in my mind right now. 1) Making it easier to write tools for testing python code, like sure[0] and freezegun[1] 2) Prototyping new features for python without touching the C code. [0] http://github.com/gabrielfalcao/sure -- where forbiddenfruit started) [1] http://github.com/spulec/freezegun