Charles Miller describes the concept of an "Anti-Test", which is a test that verifies a bug exists.
This can be written when the problem is discovered, even if it's not going to be fixed immediately.
It means that if the problem is inadvertently fixed as a result of something else, we know this has happened, have a look at why this is the case and update our records accordingly. The test can then be changed to a normal test to make sure it doesn't go wrong again.
When you're ready to fix an anti-test, you can change it to a normal test - it should now fail (and this change should be easy to make), fix the bug and make sure the test passes.
No comments:
Post a Comment