Re: [VOTE] CI tests RFC

Hi Stas On 25 April 2014 08:50, Stas Malyshev <smalyshev@sugarcrm.com> wrote: > Hi! > > I'd like to propose the CI tests RFC for the vote. Please vote at: > https://wiki.php.net/rfc/travis_ci > > The vote is planned to close on May 5. If you have any questions, please > contact me. The 4th option for the "how to handle" vote doesn't seem to gel with the other 3. It pertains mostly to what's probably a minority of contributions, where the expected behaviour for an existing test is altered - i.e. a BC break (these cases should have been thoroughly examined by the community and the failing tests should have been caught before this point, but that's another story). The other 3 describe generic course of actions that could be applied to any change that breaks the CI tests. It seems to me that this option is actually a 3rd yes/no vote, namely "should the RM be responsible for inspecting the failing tests and determining whether code is wrong or the test is wrong", so we would have: Vote 1 - general acceptance (y/n) Vote 2 - should RM be responsible for looking at failed tests (y/n) Vote 3 - preferred action when tests are failing If vote 2 has a majority "yes", then the RM would see if the test needs to be updated, and then fall down to the selected option from vote 3. If vote 2 is a majority "no", then the vote 3 result would be applied in all cases, and it would be the contributor's responsibility to find the case where the test needs updating. Thoughts? Best Regards, Chris Wright