@timofurrer thoughts?
@mle I'm fine with this, but I think it's a breaking change. I think you'd need to keep the existing flag, mark it as deprecated and then introduce the new flag to use the current behavior. I also think the suggestions from duo to make the code easier to follow make sense.
@simonschaufi Please open a support ticket so that the support team can assist you in troubleshooting what may be happening.
I'm going to close this as a support request
@dskim_gitlab Is this something we'll need to address after !207136?
cc: @francoisrose
The next iteration of the CODEOWNERS solution is to utilize DAP: DAP-Powered Intelligent Reviewer Assignment (gitlab-org#20711). I suspect the additional logic/intelligence of that solution could make auto-assignment viable for GitLab to retire roulette.
@childNode It would be easier to understand each of these as individual issues with potentially an epic for the improvements here. I think there's some things in there which are fine, and others (I'd agree with @faleksic, like count) that don't really make sense.
Also keep in mind json output is generally intended to match the equivalent API response, so we shouldn't adjust the formatting here if you're outputting json.
@mle I'll let you review, but from my side I don't think we'd introduce configuration for this nor would we want to take away the ability for people who aren't required approvers to approve. Maybe we could do something with the button states or as part of the new sidebar to emphasize it less (I actually think it already is)... but up to you on whether or not we keep this issue open.
@marc_shaw You're most familiar here - could you take a look?
@jolbrich There's already a --push flag, I think it's potentially reasonable that if you explicitly pass --push=false we could respect that as part of that process. The reason it doesn't work that way is because a branch is required to be able to create a merge request, so if the ref doesn't exist... then the whole MR create will fail.
@francoisrose setting this to backlog for now
@jcaigitlab @marc_shaw Is there any action to take here?
@francoisrose I'm putting this in the Backlog since it doesn't seem to be progressing on either side.
@slashmanov Moving to backlog for now until we prioritize this.
@dskim_gitlab I only noticed it when you click the Assign button for reviewers and use the larger flyout panel. Did you test that path?