Comments on: Python Regression tests http://www.j5live.com/2005/08/24/python-regression-tests/ Where the urethane hits the pavement Fri, 27 Aug 2010 16:17:13 -0400 http://wordpress.org/?v=2.8.4 hourly 1 By: James http://www.j5live.com/2005/08/24/python-regression-tests/comment-page-1/#comment-1664 James Thu, 25 Aug 2005 05:43:35 +0000 http://www.martianrock.com/?p=126#comment-1664 Currently in test-client.py, you have a number of tests: * connecting to the session bus * resolving an object on the session bus * narrowing to a particular interface * introspecting the object * testing various ways of invoking methods on the object * testing various inputs/outputs of methods on the object With a test suite, you usually want to run all the tests, and see which ones fail. If you stop running tests after the first error, then you won't know if the later tests pass. The docs on the Python unit test framework are here: http://www.python.org/doc/current/lib/module-unittest.html The "basic example" page shows the common way to structure the tests. Currently in test-client.py, you have a number of tests:

* connecting to the session bus

* resolving an object on the session bus

* narrowing to a particular interface

* introspecting the object

* testing various ways of invoking methods on the object

* testing various inputs/outputs of methods on the object

With a test suite, you usually want to run all the tests, and see which ones fail. If you stop running tests after the first error, then you won’t know if the later tests pass.

The docs on the Python unit test framework are here:
http://www.python.org/doc/current/lib/module-unittest.html

The “basic example” page shows the common way to structure the tests.

]]> By: J5 http://www.j5live.com/2005/08/24/python-regression-tests/comment-page-1/#comment-1663 J5 Thu, 25 Aug 2005 01:20:17 +0000 http://www.martianrock.com/?p=126#comment-1663 With make check you sort of want to stop when one test fails, fix it up and then go on. I wouldn't object however is someone sent me a patch as long as it doesn't make it any harder for people to implement tests. With make check you sort of want to stop when one test fails, fix it up and then go on. I wouldn’t object however is someone sent me a patch as long as it doesn’t make it any harder for people to implement tests.

]]> By: James http://www.j5live.com/2005/08/24/python-regression-tests/comment-page-1/#comment-1662 James Thu, 25 Aug 2005 01:13:52 +0000 http://www.martianrock.com/?p=126#comment-1662 Have you considered using the Python unit testing framework for the test-client.py implementation? It seems like it would be well suited for this kind of thing, and would continue to run tests after the first failure (i.e. it'd still fail, but you'd know if one test failed as opposed to all the tests after the failing one). Have you considered using the Python unit testing framework for the test-client.py implementation? It seems like it would be well suited for this kind of thing, and would continue to run tests after the first failure (i.e. it’d still fail, but you’d know if one test failed as opposed to all the tests after the failing one).

]]>