[llvm] [DevPolicy] Add guidance to not submit code or reviews on someone else's (PR #69701)

[Sam McCall via llvm-commits]

sam-mccall wrote:

Per https://discourse.llvm.org/t/submitting-code-or-review-comments-on-someone-elses-behalf/74269, this change "helps with not accidentally helping a banned person to circumvent their ban", and this PR is the right venue to for questions/feedback about this.

The developer policy doesn't speak about bans, but the code of conduct says about bans:

> We are not interested in evaluating severity, responding punitively, or holding people accountable. Both the relevance and our response is instead focused on how a person’s continued participation impacts the community’s safety, wellbeing, and inclusivity. We specifically prioritize remaining a welcoming community to victims as well as groups subjected to systemic marginalization or underrepresentation.

Assuming "on-behalf-of" contributions themselves meet the bar we set in terms of technical content and respectful conduct, and they are not attributed to the banned individual, it's not clear to me what the impact of inclusivity. Avoiding circumvention of bans as a goal *per se* a seems more like "holding people accountable".

I'm speculating here with limited perspective. From the announcement it's unclear whether this policy change is motivated by reports of harm, or on equally-shaky speculation! More context would help.

tl;dr: https://xkcd.com/810/?


https://github.com/llvm/llvm-project/pull/69701