Hi everyone! After some discussion with both upstreams and the QA team, we came up with a few modifications to the landing process and image promotion criteria. The process will now become a little bit more strict, but we tried finding a middle ground between ensuring high image quality and landing convenience.
What changes will be happening? First of all, whenever any test failure gets spotted by CI or the Landing Team, a manual reproduction of the test steps will be performed to check if the failure is a real regression or a test/infrastructure regression. Depending on the result of that, a bug will be filled and assigned either to QA or the upstream developers. Now, if the test failure is a regression the standard thing happens - QA decide if this blocks promotion or not. In both cases a 'timer' will be started for the given issue. This timer will be increased by one every day the issue remains unfixed. Once it reaches a certain value, the issue becomes a promotion blocker by rule and also blocking non-fix landings for the specific component. The timer value for becoming a blocker will, for now, be set to 7 days. This means upstream developers (or QA engineers) will have 7 days to fix the issue before it becomes a bother to everyone. This value will be tweaked after initial feedback of how well the 'system' works. All necessary information will be made available on the daily landing team messages. We can think about other ideas of visualization later on. The result we want to achieve is putting a little bit pressure on upstreams not to leave smaller issues 'lying around' on the image because of their low priority. Leaving them unattended for a few days is ok, but we had some that have been around since even months! Let's try it out! See you next week! Best regards, -- Łukasz 'sil2100' Zemczak [email protected] www.canonical.com -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : [email protected] Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
