On Mar 24, 2012, at 5:32 PM, Nicolas Cellier wrote:

Then if I introduce a new test documenting an old bug, it will then be rejected right ?

No, it will be added to the bug tracker.

I wonder if I'd really want such a dumb system without an ounce of human expertise...

In the past we did exactly that: add failing tests, so we will forced to fix them! The thing is: it did not work. Nobody cares about fixing failing tests. The only result is that the build is in the state "some tests are failing". A week later "some more tests are failing". A month later "When did that test start to fail ?!". In addition: how do we know that a vm build is good? That we are in a good state? Marcus -- Marcus Denker -- http://marcusdenker.de