On Wed, Aug 10, 2016 at 6:47 AM, Lars Bergstrom <larsb...@mozilla.com> wrote:
> But in this model, there's also "is this a failure from one of the N > PRs before me?" which isn't a type of failure that is common (AFAIK) > when contributing to OSS projects. Intermittents, especially when they > lead to PR bitrot, already chase off users in Servo. I'm worried this > extra failure mode will compound that problem. > Strongly agree -- this can be a huge pain point. When I was doing rollups in Rust this used to be a major hurdle in getting the rollups merged; finding out which of the ~10 PRs caused a failure is not always straightforward and there's often ambiguity. This was still fine because I was the only one investigating these failures, but part of this effort will be multiplied by 10 if there is no "sheriff" handling the rollups. Intermittents make the problem worse, and we have many of those. Note that we don't lose pre-commit testing in the proposed autoland model. Our cycle times become 3 hours, though. There is a (expectedly rare) failure mode where stylo-integration fails to merge cleanly into m-c; which is a natural byproduct of trying to marry two opposite CI models together. But like I said it's expected to be rare so it shouldn't be an issue. Thanks, -Manish Goregaokar _______________________________________________ dev-servo mailing list dev-servo@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-servo
