Comments on: Unit tests don’t guarantee that your system works http://blog.codevader.com/2008/03/05/unit-tests-dont-guarantee-that-your-system-works/ Thoughts from the nice people at Codevader :) Thu, 07 Aug 2008 20:03:24 +0000 http://wordpress.org/?v=2.2.2 By: Maurício http://blog.codevader.com/2008/03/05/unit-tests-dont-guarantee-that-your-system-works/#comment-58 Maurício Fri, 18 Apr 2008 14:18:22 +0000 http://blog.codevader.com/2008/03/05/unit-tests-dont-guarantee-that-your-system-works/#comment-58 Hi Mark, The "validates_uniqueness_of" is just an example of "integrated" test. Everytime your system needs to "speak" to an outside entity there should be integration tests asserting that they are talking correctly and you never know if your SQL code works without running it on the "real" database. Hi Mark,

The “validates_uniqueness_of” is just an example of “integrated” test. Everytime your system needs to “speak” to an outside entity there should be integration tests asserting that they are talking correctly and you never know if your SQL code works without running it on the “real” database.

]]> By: Mark Wilden http://blog.codevader.com/2008/03/05/unit-tests-dont-guarantee-that-your-system-works/#comment-41 Mark Wilden Thu, 06 Mar 2008 15:24:49 +0000 http://blog.codevader.com/2008/03/05/unit-tests-dont-guarantee-that-your-system-works/#comment-41 It seems to me that the only reason cited for not using BDD for testing validates_uniqueness_of is that it's "overkill" to mock the connection. Without that one word, the conclusions of an interesting and thoughtful post are unsupported, in my opinion. What if we decided that mocking the connection was not "overkill"? Has anyone done it? I haven't, but I've often thought about doing it. This article makes me want to explore the idea more. One more point: The purpose of programmer tests is not to ensure correctness. A great programmer will never be able to do the work of a great QAer, because each have diametrically opposed goals. The programmer wants his code to work - the QAer wants to make it fail. Just some thoughts. I found the rationale for TDD/BDD presented here very clear, and I plan to use it when explaining their benefits. ///ark It seems to me that the only reason cited for not using BDD for testing validates_uniqueness_of is that it’s “overkill” to mock the connection. Without that one word, the conclusions of an interesting and thoughtful post are unsupported, in my opinion.

What if we decided that mocking the connection was not “overkill”? Has anyone done it? I haven’t, but I’ve often thought about doing it. This article makes me want to explore the idea more.

One more point: The purpose of programmer tests is not to ensure correctness. A great programmer will never be able to do the work of a great QAer, because each have diametrically opposed goals. The programmer wants his code to work - the QAer wants to make it fail.

Just some thoughts. I found the rationale for TDD/BDD presented here very clear, and I plan to use it when explaining their benefits.

///ark

]]> By: Derek http://blog.codevader.com/2008/03/05/unit-tests-dont-guarantee-that-your-system-works/#comment-40 Derek Thu, 06 Mar 2008 15:12:03 +0000 http://blog.codevader.com/2008/03/05/unit-tests-dont-guarantee-that-your-system-works/#comment-40 Another method is to load your model and test that it executes the validates_uniqueness_of method with the expected parameters. Another method is to load your model and test that it executes the validates_uniqueness_of method with the expected parameters.

]]> By: Unit tests don’t guarantee that your system works « Codeshooter’s Weblog http://blog.codevader.com/2008/03/05/unit-tests-dont-guarantee-that-your-system-works/#comment-39 Unit tests don’t guarantee that your system works « Codeshooter’s Weblog Wed, 05 Mar 2008 15:54:02 +0000 http://blog.codevader.com/2008/03/05/unit-tests-dont-guarantee-that-your-system-works/#comment-39 [...] Originally published at the CodeVader blog [...] […] Originally published at the CodeVader blog […]

]]>