On Sat, Apr 18, 2009 at 06:20:48PM -0700, Istvan Albert wrote: -> On Apr 18, 12:09?am, "C. Titus Brown" <c...@msu.edu> wrote: -> -> > If people are ok with these changes, the next step would be to remove -> > the various get_suite functionality from the test files. -> -> I'm fine with the change. One possible issue with it (if I read -> correclty) is that -> one loses the ability to choose the order in which tests are run. -> -> This can be handy in several circumstances. First as convenience, when -> adding a new test to the suite, you can list it first, thus saves you -> time as it fails first. But beyond that occasionally I have had the -> need to prepare the input in one set of tests, then work off of these -> results in other sets of tests. While never explicitly stated, these -> were part of the rationale that went into the choosing to implement -> the get_suite() function.
OK -- but why wouldn't we just automate that dependency calculation anyway? Time/space issues? Generally it's thought to be a bad idea for unit and functional tests to NOT be order independent (and in fact "shuffling" the test order can discover hidden caching and layer bugs quite easily). I'll take your e-mail as a +0 in any case ;). Anyone else have any comments? cheers, --titus -- C. Titus Brown, c...@msu.edu --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "pygr-dev" group. To post to this group, send email to pygr-dev@googlegroups.com To unsubscribe from this group, send email to pygr-dev+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/pygr-dev?hl=en -~----------~----~----~----~------~----~------~--~---
