On Sat, Feb 05, 2005 at 07:21:01PM -0400, bulia byak wrote:
On Sat, 05 Feb 2005 18:00:14 -0500, MenTaLguY <mental@...3...> wrote:
On Sat, 2005-02-05 at 15:34, bulia byak wrote:
I don't feel this is safe enough. You change not only rollback but also commit. At the very least, I want Mental to OK these changes. And in general, I would prefer a release with some failed tests than with broken core.
Unless the tests are really wrong, failed tests mean broken core by definition...
Quite often, failed tests mean that they are not updated when the core is changed.
Additional information relevant to the prior-to-review assessment of probability that the test failure indicates a real problem:
In this case, the test triggered an assertion in core. Tests that trigger assertions in core are more likely to indicate a true problem than other test failures.
(Just a probabilistic argument, I haven't reviewed the change in this particular case.)
pjrm.