[mgunit] mgunit 1.1 release coming up

[Michael Galloy]

On Feb 5, 2010, at 1:13 PM, Symons, Greg - GS wrote:

> Glad to hear that 1.1 is going to be released soon! I've been using trunk for some new development and it seems pretty good so far. I was really excited to discover the junit format... I'd been thinking of doing that work myself so that I could get Hudson to report unit-test results for continuous integration.

Yes, someone else requested the JUnit output to use with Hudson. I guess I should learn more about Hudson.

> I didn't notice the skipped stuff. Is that for the whole test case, or can it be set on an individual test method basis?

Individual test cases can be skipped. What I have been doing is adding a line in a test something like:

  assert, (some_condition) || self.skipped++, 'message why test was skipped'

Then, if "some_condition" were not true, then self.skipped will be set and the given message will be shown for why the test was skipped. I might introduce a keyword to ASSERT for this case.

I have been using skipped tests like this in GPULib, where the double precision tests do not make any sense unless running on a graphics card that is capable of doing double-precision calculations (not all are). This way, my tests results make a bit more sense and I don't have to constantly be telling people "it's OK that 50 tests fail, your card can't do double-precision".

> I haven't tried the new assert message stuff yet; I'd already written my own wrappers around assert to get something closer to what you get from JUnit's assertEqual et. al. But I may give it a try if I get some time to rework those wrappers.

Mike
--
www.michaelgalloy.com