On Mon, Apr 19, 2010 at 9:48 AM, Marius Gedminas <marius@gedmin.as> wrote:
On Mon, Apr 19, 2010 at 12:50:23PM +0200, Adam GROSZER wrote:
I'm somewhat vary on unittests. I've seen some damn cryptic ones that took a lot of time to decipher. A doctest somehow forces you to dump your mind (well at least that, if we're not that brilliant techdoc writers).
That's true, but if the doctest gets too long, any readability advantages are negated.
If you've the discipline to keep the doctests short, I don't see why you shouldn't continue writing them instead of unit tests -- and by "short" I mean 1-7 statements:
You might be interested in Manuel's isolation mechanism: http://packages.python.org/manuel/#test-isolation
The downside of this style is that it's difficult to refactor common bits from the doctests into shared methods, so you end up with a lot of duplicated code.
If I understand you (which I'm not sure I do), I prefer to factor out common code into test helpers that are either referenced in footnotes (to keep all the code in one place) or in a separate module. -- Benji York