On Tue, 2010-06-15 at 11:37 -0700, Adam Williamson wrote: > On Tue, 2010-06-15 at 17:23 +0000, "Jóhann B. Guðmundsson" wrote: > > On 06/15/2010 04:08 PM, Adam Williamson wrote: > > >> We can't test manually every possible feature. > > >> > > > Yeah, this is how I generally handle this kind of case. We can't have > > > ten thousand test cases for every possible configuration wiggle. > > > > > > > Ok. so let me ask both of you this. > > > > What's the criteria for us to use when choosing which option we use in > > test cases when we write them? > > In general, everything should be default or the most 'obvious' / > 'common' choice. That's the rule of thumb I go with. I agree with Adam. Those are definitely the guidelines I wish to follow. Unfortunately, if you were to look at tests I've written/touched in the past, you'll probably find cases where I've diverged from this approach. Just like a fine wine, our collective process matures over time :) I've been thinking about this in response to some of the install test matrix cleanup Hurry has been doing. One approach I used a lot in different test circles is to define "variations" of the test. Similar to the changes Jóhann lists earlier, test variations were used to describe _minor_ changes to the test instructions/results. If desired/useful, I can add an optional "Variations" section to the test case template. [1] Thanks, James [1] https://fedoraproject.org/wiki/Template:QA/Test_Case
Attachment:
signature.asc
Description: This is a digitally signed message part
-- test mailing list test@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
