How do I diagnose test errors?

Peter Arrenbrecht peter.arrenbrecht at gmail.com
Thu Jun 5 04:22:46 CDT 2008


On Thu, Jun 5, 2008 at 11:10 AM, Peter Arrenbrecht
<peter.arrenbrecht at gmail.com> wrote:
> On Thu, Jun 5, 2008 at 10:45 AM, Martin Geisler <mg at daimi.au.dk> wrote:
>> "Peter Arrenbrecht" <peter.arrenbrecht at gmail.com> writes:
>>
>>> Having a way to cite actual tested input and output in documentation
>>> would be great! (See http://arrenbrecht.ch/jcite/ for my take on
>>> this for Java.)
>>
>> I looked at the examples on your page -- is this not similar to the
>> standard doctest module in Python?
>>
>>  http://docs.python.org/lib/module-doctest.html
>
> Yes, I believe so, although I don't know how doctest handles setup
> code you don't want to cite, highlighting, etc. And AFAIK doctest does
> not alert you to changed output (meaning you should inspect
> surrounding text for whether it is still accurate). Also, most Hg
> tests are not Python- but script-based.

Follow-up: Take, for example, this page from an actual project using
JCite extensively:

  http://www.formulacompiler.org/quick.htm

While the code _is_ cited from a test, the citations focus on how to
do things, not on sometimes distracting verification code. It's like
literate testing. (Of course, Python tends to make things less
distracting, so doctest might be totally sufficient there.)
-parren


More information about the Mercurial-devel mailing list