Rana Dasgupta wrote: > On 6/26/06, Geir Magnusson Jr <[EMAIL PROTECTED]> wrote: >> >> >Odd. It was broken for me too. What changed? >> >> >I applied the patch - thanks. > > > To determine this, one would have to roll back the last commit and retry > the build + smoke tests. And traverse back till we identify what broke it. > We could also consider requesting the patches that came after this to be > resubmitted. At least. > It may help to start putting in some submission citeria for patches eg., a > successful log from the existing drlvm smoke tests etc. Patches, though not > big, are coming in quite rapidly and it can't be easy for committers to > verify everything on two target platforms. > Also, the existing few smoke tests will not catch all bugs. At some point, > new tests should accompany non trivial submissions. We don't want to get > into a situation where we have a serious breakage. > It will be good to know how we are doing the corresponding stuff in the > classlibs.
The short answer is: by relying on the test suite. I agree that the first goal is to reach a stable state, then ensure that patches come with new tests maintaining the stable state. Regards, Tim -- Tim Ellison ([EMAIL PROTECTED]) IBM Java technology centre, UK. --------------------------------------------------------------------- Terms of use : http://incubator.apache.org/harmony/mailing.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]