On Sun, May 22, 2011 at 01:26:36AM -0400, Edward Cherlin wrote: > At the end of your e-mail, you [Daniel] wrote > > > Testing is much appreciated. > > It sounds as if there is no formal testing process for Sugar on XOs.
Not quite sure what you mean by formal. Some of the planned and funded testing has been described by Martin and Samuel. We also hope and expect the community here on devel@ and testing@ to do their own testing and report their findings. That's one of the reasons why Daniel and other releasers call for testing publically. Pointing to the test plans in a release would unnecessarily constrain the testing done by the community. Random discovery may exceed the coverage of a test plan. So the absence of a test plan shouldn't be construed as meaning there is no formal testing. The testing done by OLPC, devel@ and testing@ should and does reach the developers. Other distributions will pick up fixes gradually, or if they focus their efforts they can pick up fixes early. It would be easier if Sugar and activities could be tested in a semiautomatic manner. There's no built-in tests in Sugar and activities, but that's an upstream problem that Sugar Labs have occasionally had people look into. > I would like to see whether I can help with this situation, by > recruiting more testers, and setting up something more formal to make > sure that issues don't fall through cracks. Engaging in the existing processes might involve: - giving the smoke test plans to testers, - asking testers to test fixes in a release, by reading the test plans in a ticket, or the ticket problem description, - asking testers to randomly test releases, according to their own expectations, - training testers in the search and report techniques, http://wiki.laptop.org/go/Reporting_bugs#Harder_but_more_helpful:_Trac_with_care On Sun, May 22, 2011 at 09:02:48PM -0400, Samuel Greenfeld wrote: > * The two main organizations within OLPC each have a test/QA engineer > helping them. James Cameron tends to focus more on the low-level > hardware; I tend to focus more on the higher-level software. Yes. But we intentionally tread on each other's turf to get the job done. If I spot an interesting Sugar or activity bug, I do report it. > * In general though I agree we need a better way to split the load > if other formal test volunteers are available. To split the load effectively takes one more step ... test PASS reporting. If we know that n other people have done a test without problems, it means we need not concentrate on doing it. So often we only get the test FAIL reporting. So if a volunteer tests something, and it works, that is just as useful for publication as a test failure report. I'd like to encourage devel@ subscribers who test builds to actually say they did so, regardless of whether they had anything more to say. -- James Cameron http://quozl.linux.org.au/ _______________________________________________ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel