Comments on: TDD: Testing delegation http://www.markhneedham.com/blog/2009/11/27/tdd-testing-delegation/ Thoughts on Software Development Sat, 31 Jul 2010 12:46:58 +0000 hourly 1 http://wordpress.org/?v=3.0 By: Literal Laws « I Built His Cage http://www.markhneedham.com/blog/2009/11/27/tdd-testing-delegation/comment-page-1/#comment-31060 Literal Laws « I Built His Cage Thu, 28 Jan 2010 03:30:38 +0000 http://www.markhneedham.com/blog/?p=1879#comment-31060 [...] Mark Needham has blogged in the past about using TDD on delegation. However, the post is mostly about unit-testing such objects, not writing the tests first in order [...] [...] Mark Needham has blogged in the past about using TDD on delegation. However, the post is mostly about unit-testing such objects, not writing the tests first in order [...]

]]> By: Matt Dunn http://www.markhneedham.com/blog/2009/11/27/tdd-testing-delegation/comment-page-1/#comment-27354 Matt Dunn Sun, 29 Nov 2009 02:46:29 +0000 http://www.markhneedham.com/blog/?p=1879#comment-27354 Similarly to what you're saying, I've found interaction-based testing provides the most value in defining/describing an interaction when mocking explicit roles (and hasn't been coded yet). So using mocking to discover roles rather than substitute dependencies is where the real value lies. In light of this in your examples that use partial mocking it makes sense to use a concrete stub instead. I'm also starting to think concrete stubs are the way to go as real code is easier to keep up to date than some of the tricks that various testing frameworks use. It depends on how powerful your tools are at to keeping up with changes though. Heres a more extreme view to mull over: http://www.pubbitch.org/blog/2006/09/09/rules_of_engagement_for_using_mock_objects Similarly to what you're saying, I've found interaction-based testing provides the most value in defining/describing an interaction when mocking explicit roles (and hasn't been coded yet). So using mocking to discover roles rather than substitute dependencies is where the real value lies. In light of this in your examples that use partial mocking it makes sense to use a concrete stub instead.

I'm also starting to think concrete stubs are the way to go as real code is easier to keep up to date than some of the tricks that various testing frameworks use. It depends on how powerful your tools are at to keeping up with changes though. Heres a more extreme view to mull over:

http://www.pubbitch.org/blog/2006/09/09/rules_of_engagement_for_using_mock_objects

]]> By: Reflective Perspective - Chris Alcock » The Morning Brew #486 http://www.markhneedham.com/blog/2009/11/27/tdd-testing-delegation/comment-page-1/#comment-27297 Reflective Perspective - Chris Alcock » The Morning Brew #486 Fri, 27 Nov 2009 08:22:54 +0000 http://www.markhneedham.com/blog/?p=1879#comment-27297 [...] TDD: Testing delegation - Mark Needham talks about the evolution of his testing style when testing delegation based code, looking at how he now perfers stubs to full mocks in this situation [...] [...] TDD: Testing delegation – Mark Needham talks about the evolution of his testing style when testing delegation based code, looking at how he now perfers stubs to full mocks in this situation [...]

]]>