I've been using doctest for a while now, and I like it most of the time. It's annoying when I expect a number that has to fall within a certain range or a string that has to meet certain conditions. In this case, the recommended approach seems to be to encode the condition as a logical operator returning True or False, e.g., """ >>> from math import pi >>> 3.14 < pi < 3.15 True """ But this approach leads to a lot of True lines in my doctests. And if a test fails, I get a message that says something like, "Expected True; got False." But I just discovered that I can use assert to accomplish the same thing: """ >>> from math import pi >>> assert 3.14 < pi < 3.15 """ The input is more compact. And this way, if the assertion fails, doctest says, "Expected nothing; got AssertionError..." followed by the offending assert statement, which is more readable. Is there any reason not to do this? Continue reading...
