Re: [Potential Spoof] Re: [presto-tsc] Nominating Leiqing as new committer
rongrong100@...
> I expect committers to have substantial expertise on a subset of core parts of the engine. The verifier scope is quite limited and having committer with expertise in just that area is of limited value. I mostly agree with this, which is why when James proposed to add Leiqing earlier I pointed out that he mostly only worked on verifier, which might not be of enough scope. I changed my mind and decided to nominate him because I think there's still great value in this. Leiqing is the expert on verifier and the new benchmark work is very much related to verifier. His code review is of high quality and I rarely found myself needing to add additional comments. So at least in these sub-systems, he "had a track record of providing quality reviews and approving PRs such that existing committers don’t need to review the code and can trust the to-be-committers approval". If we have an area committership model I'd nominate him as the area committer for these limited scopes. I don't see any value in having to insert me, or any other committer in these reviews to merge code. None of the committers know every subsystems in the code base, and we trust them to say "I'm not the right person to review, please ask xxx instead". And I trust Leiqing to do the same on other subjects he's not familiar with. > Furthermore, I expect to-be-committers to have a track record of providing quality reviews and approving PRs such that existing committers don’t need to review the code and can trust the to-be-committers approval. In other words, I’d expect to see existing committers delegating code reviews to to-be-committers on a regular basis. I don’t see evidence of that. I agree with this point. Except that I saw evidence. When you said you "don't see evidence of that", do you mean that you haven't worked with him thus haven't seen his work, or that you've seen evidence that he didn't meet the bar? Overall what I want to achieve with this nomination is the experts in the subsystems could have some autonomy on the subsystems they build. If committer is not the right way to do it, please propose different solutions. I don't want to be the artificial gate keeper because I don't see myself adding any value to the process. If someone else want to volunteer for doing all final reviews on verifier / benchmark runner, please go ahead. Thanks! Rongrong
On Tue, Nov 19, 2019 at 4:01 AM Masha Basmanova <mbasmanova@...> wrote:
|
|