More failing tests...

[William Reade]

On Thu, 2013-02-28 at 20:01 +1000, Ian Booth wrote:
> > 
> >  1) Do *not* merge into trunk without a full passing test suite.
> >     Frankly, you shouldn't even lbox propose without the justified
> >     belief that you haven't broken anything, and the only way to get
> >     that is to RUN THE TESTS.
> >
> 
> Easier said then done with the false positives. I *did* run the full test suite
> and got one failure, which I was pretty sure was one of the intermittent ones. I
> may have been wrong. So long as we have intermittent test failures, we *will*
> have this same problem over again, since someone may fluke a clean test run and
> the next guy may get the failure.

If you see a failure and believe it's intermittent, please run it again
until you're sure it is, and make sure there's a bug for it. In this
instance there were 9 reliably failing tests in worker/uniter, and I'm
not sure how one would mistake one for the other. You *are* merging
trunk into your branch before you run the tests and propose/submit,
right?

If you're lucky enough to be able to reproduce an intermittent failure
at a decent rate, please step up and try to actually fix it: we all have
subtle differences in hardware/OS, and if you're suffering excessively
from one particular failure you're probably the best person to deal with
it. If the issue is unclear, ofc, please seek guidance from the most
likely responsible individual.

Reasonable?

Cheers
William