Rasjid Wilcox <ras...@familywilcox.net> writes: > To date I've mostly used Nose for my unit testing, but was just > wanting to canvas views on what the current state of the art in Python > testing is, and if I should be looking at something else.
I use Python 2.7, or Python 3, and use the standard library ‘unittest’ along with ‘testtools’ <URL:http://testtools.readthedocs.org/> and ‘testscenarios’ <URL:https://pypi.python.org/pypi/testscenarios>. The ‘unittest’ library in Python's standard library has been significantly improved. Developers using Python 3 get the full benefits, but Python 2.7 also had many of the improvements back-ported. * Test case discovery * Specify test cases from command line * More assertion methods, more comparison methods * ‘assertRaises’ as a context manager (for ‘with’) * Adding cleanup functions * Skipping tests conditionally * Class-level and module-level fixtures * etc. Michael Foord was a primary developer and advocate of these improvements. Here is his description from 2011 <URL:http://www.voidspace.org.uk/python/articles/unittest2.shtml>. All those improvements, along with the wonderful author-defined Matchers in ‘testtools’ and the data-driven testing from ‘testscenarios’, I virtually have no need of anything more than those. And they're easily accepted by my team, whereas switching to a different test framework would be (correctly) met with much more resistance. -- \ “Sunday: A day given over by Americans to wishing that they | `\ themselves were dead and in Heaven, and that their neighbors | _o__) were dead and in Hell.” —Henry L. Mencken | Ben Finney _______________________________________________ melbourne-pug mailing list melbourne-pug@python.org http://mail.python.org/mailman/listinfo/melbourne-pug
