<div class="gmail_quote">On 12 August 2010 13:30, James Cameron <span dir="ltr"><<a href="mailto:quozl@laptop.org">quozl@laptop.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"> <div class="im">> 2) Which testing framework to use</div><div class="im"> <br> </div>Whatever the test developer and activity developer can agree on.<br> <br> I don't know enough about the options to choose.<br> <br> (p.s. I'm aware of your work on Sugarbot. While this is fantastic, I'm<br> worried that a test that depends on Sugarbot might end up not being used<br> because Sugarbot is unavailable on the platform, or a later version is<br> needed.)</blockquote><div><br></div><div>Oh, I was mainly talking about unit testing. Something like Sugarbot / dogtail is more suited to larger-scale testing.</div><div><br></div><div>I see the choice here between derivatives of Kent Beck's SUnit and some of the newer testing frameworks. In Python, this would be the unittest* module. The other two options I listed were nose [1] & py.test[2] tests. There are other possibilities, such as zope.testing.</div> <div><br></div><div>In my opinion, nose & py.test make writing tests much simpler. I personally prefer py.test because of some of its additional capabilities, such as being able to send tests to other cores/machines via ssh with a command line switch. I think nose has a greater popularity primarily because it has a better website with friendlier documentation.</div> <div><br></div><div>Tim</div><div><br></div><div>* From Py 2.7, that is depreciated in favour of unittest2.</div><div>[1] <a href="http://somethingaboutorange.com/mrl/projects/nose/0.11.2/">http://somethingaboutorange.com/mrl/projects/nose/0.11.2/</a></div> <div>[2] <a href="http://codespeak.net/py/dist/test/">http://codespeak.net/py/dist/test/</a></div></div>