Hi folks, I've moved 2.0 build-and-test's branch to the trunk - the branch was used during M2-M5 testing and IMO it proved to be very useful (also I tagged the former trunk as 1.0.)
I think that this is a good time to collect ideas and hear opinions what should be improved and possible directions for test infra development. I have to admit that the current testing infra is far from being perfect – there a lot for work to make is more usable, cleaning up current scripts, sort out all know issues with tests/scenarios and so on. But I'd like to address problems I believe of different level. I have an impression that Harmony tests suites are still in shadow (IOW, IMHO they are undervalued) of testing automation (i.e. CruiseControl and support scripts) – and most of folks prefer to contribute to the code development. And I believe that it is possible to make tests development also interesting too. IMO the missed point that most of Harmony test suites can be used as stand alone suites to verify/test different implementations – in other words, the suites are not tightly coupled with DRL VM, for example, it is possible to run them against IBM VME (BTW, may it be makes sense to run full set of suites against IBM VME for each milestone candidate as for DRL VM. Any volunteers?) I'm going to update the test infra home page [1] with info related to the renewed trunk. Also I think the page should contain some words about project's testing strategy: i.e. how Harmony QA and release process are organized - integrity/snapshot testing, milestones, code freeze and etc., what suites are there (their description and purpose). I think this may encourage people to invest efforts into developing tests. Other ideas and thoughts? [1] http://harmony.apache.org/subcomponents/buildtest/index.html Thanks, Stepan.
