[PATCH] D99305: [docs] Document our norms around reverts

[Philip Reames via Phabricator via llvm-commits]

reames added inline comments.


================
Comment at: llvm/docs/DeveloperPolicy.rst:336-337
+  opportunity enables this.
+* The other case for a third-party revert is for serious norm violations.
+  Such situations are rare.
+
----------------
rupprecht wrote:
> I don't know if I have a specific suggestion that is better, but I'm not a fan of the wording here, especially with the "serious norm violations" wording.
> 
> While I agree with the general policy around timely reverts, starting a discussion before reverting someone else's change, etc., I also think that documenting these as "serious norm violations" unnecessarily escalates these cases when they do happen. If someone has reverted a change that falls into this category, there's a decent chance that it wasn't an intentional norms violation, and the discussion should start out with "Hey, can we talk more about this revert? I don't know if reverting was the right decision here" and not "This is a serious norms violation as documented in [link to these docs], please undo the revert".
I want to avoid getting into the discussion of norm violation handling in this review.  You raise several good points, and we should eventually document this, but I expect it to be an involved discussion.  I really want the basics landed before opening that can of worms.

If anyone can think of wording to further soften this sentence, please throw it out.  I can't think of anything without expanding this bullet a lot, and every attempt I've made ran into something I thought needed discussion I'm hoping to avoid on this review.  

Maybe I should just strike that bullet for now entirely?  The previous bullet reads reasonably on it's own as an answer to the question.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D99305/new/

https://reviews.llvm.org/D99305