Begin forwarded message:
From: Keith Hodges <keith_hodges@yahoo.co.uk> Date: November 8, 2008 2:18:32 AM CEST To: Alexandre Bergel <Alexandre.Bergel@inria.fr> Cc: stephane ducasse <stephane.ducasse@free.fr> Subject: Re: [Pharo-project] Sunit Extensions breaking the TestRunner
Alexandre Bergel wrote:
The situation is getting messy.
Cheers, Alexandre exactly... everyone doing their own thing for their own reasons. Will you be willing to help restore some coherence to the picture.
I set a few goals that are fulfilled in my derived implementation. My improvements are based upon more than 10 years of experience of using SUnit and things like it. These have been out there for 2 years, so there is no reason why people cannot build upon this base.
1. Ability to define test suites, and have a UI for running test suites. I have been using this facility for 8 years in st/x and always miss it in squeak. I want to be able to define a suite for "My Project".
2. Ability to flag which tests are expected to work in which situation, which image version and which platform. So that I can run with "all tests green" even if I happen to have a few windows specific tests, when I am running on a mac.
3. Ability to use tests to specify behaviour and requirements even before the tests work. I.e. mark tests as #toDo.
4. Ability to run tests in an automated tool, without a UI.
5. Ability to sort tests so that they are run fastest first. 95% of the tests will run in 10% of the time.
6. Ability to determine whether a test used the network or not, and filter out those tests if necessary.
7. Ability to integrate other testing technologies like SSpec. (not actually done, but conceptually possible now)
A couple of other improvements are also included, such as the ability to define a custom exception handler around all of your tests.
In my opinion some members of the Pharo team have a big problem with "Not Invented Here". They don't even look at the existing stuff before diving in.
I spent several months integrating the fixes from 3 separate streams of Monticello, specifically with the goal of promoting coherence within all of the the squeak communities. I did this work for the benefit of all squeak users, not for the benefit of my health or financial well being. In fact it has cost me several thousands of pounds in lost revenue.
I am actually fairly insulted by the attitude of the Pharo team towards this work. Since they take MC and continue to work on it separately, it actually renders all of my effort a waste of time.
perhaps we can rescue sunit before it is too late.
Keith