PyJUnitXML, a pyunit extension to output JUnit compatible XML.
A Python unittest TestResult that outputs JUnit compatible XML.
PyJUnitXML is Copyright (c) 2009 Robert Collins, and distributed under the GNU Library General Public Licence version 3 (LGPL-3). Broadly, this means you have the right to use this library in any application you want, regardless of the licence of your application, but that if you modify this library the modifications you make need to be released under a compatible terms if you convey the application to other users. For exact details see the COPYING and GPL-3 files included in the source.
To use PyJUnitXML you need to specify it as your TestResult when running your tests using pyunit. One way of doing this is to invoke the test method directly on your test suite:
>>> import junitxml >>> result = junitxml.JUnitXmlResult(sys.stdout) >>> result.startTestRun() >>> suite.run(result) >>> result.stopTestRun()
The startTestRun and stopTestRun methods are new in python 2.7 and 3.1. If you are using older versions of python, and do not invoke the test result directly, you may find that no output occurs at all. This is likely because the stopTestRun method is not being invoked, and you will need to arrange for it to be invoked.
There is a command line runner you can use:
$ python -m junitxml.main discover
$ junitxml discover
If the package scripts have been installed.
My test runs, but nothing is output? Check that result.stopTestRun is being called. Due to the XML file format requiring a header containing the test counts, nothing is output until the test run has completed, and that is signalled by the stopTestRun method being called.
Reporting Bugs and Asking Questions
PyJUnitXML is hosted on Launchpad at https://launchpad.net/pyjunitxml. Please file bugs, ask questions and so on at will.
Running PyJUnitXML’s test suite
The function junitxml.test_suite defines the junitxml test suite, you can use any runner you like. I personally use:
$ python -m subunit.run junitxml.test_suite | subunit2gtk
If you have testrepository you can simply run testr:
$ testr run
Enjoy, Rob Collins