By the way, since we discussed Copilot review a few days ago, I will note that Copilot made two review comments on this particular PR, and neither of them were correct. (The arguments to UDFs are never null, because of strictness, and ByteString is a Calcite class so can be used in doc without qualification.) Fixing them would have made the code worse.
> On Apr 1, 2026, at 8:17 AM, Cancai Cai <[email protected]> wrote: > > Oh, thanks for reminding me. It seems I missed that discussion (I wasn't > involved in Calcite back then). > > Then let's continue with this approach. Thank you for your patient > explanation. > > Best wishes > Cancai > >> On 2026/04/01 15:05:46 Stamatis Zampetakis wrote: >> The person who merged PR#4855 is a calcite committer so it is possible >> and accepted to merge a PR without getting an explicit approval by >> another committer. >> >> There are some past discussions on the topic [1, 2] that basically sum up to: >> * ask and wait for review when changes are big, important, breaking (RTC) >> * push small and low risk changes without waiting for a review (CTR) >> >> This model has been working well for many years so I don't find it >> necessary to change it. >> >> Best, >> Stamatis >> >> [1] https://lists.apache.org/thread/fsrv3hdck20374y1c9s380h5cxhk6gnf >> [2] https://lists.apache.org/thread/x61yz4j4d5820dgbrtynmkzos96x4k36 >> >>> On Wed, Apr 1, 2026 at 4:47 PM Cancai Cai <[email protected]> wrote: >>> >>> Hi, Stamatis >>> >>> Okay, perhaps I've been a bit vague. Let me illustrate with an example, >>> such as a recent merged PR: >>> >>> https://github.com/apache/calcite/pull/4855 >>> >>> This PR didn't have any other committers or PMCs approving it. >>> >>> Best wishes, >>> Cancai. >>> >>> On 2026/04/01 14:40:08 Stamatis Zampetakis wrote: >>>> Hi Cancai, >>>> >>>> Can you please provide some pointers to the specific PRs? Only >>>> committers have write access to the repo so not sure what exactly is >>>> the problem you are referring to. >>>> Note that in Calcite we don't have a strict RTC policy and for many >>>> cases we have been doing CTR. >>>> >>>> Best, >>>> Stamatis >>>> >>>> On Wed, Apr 1, 2026 at 4:38 PM jensen <[email protected]> wrote: >>>>> >>>>> Thanks for raising this topic. I think establishing a basic principle for >>>>> the Apache Calcite community is necessary. Having multiple reviewers can >>>>> help contributors double-check their code, which is a good practice. As >>>>> you mentioned, all of Calcite’s contributors are volunteers, and their >>>>> time is limited. I believe contributors should be patient throughout the >>>>> contribution process. I’m not an expert in all areas of Calcite, but I’ll >>>>> continue learning and try to review more pull requests. >>>>> >>>>> >>>>> >>>>> Best regards, >>>>> >>>>> Zhen >>>>> >>>>> ---- Replied Message ---- >>>>> | From | Cancai Cai<[email protected]> | >>>>> | Date | 04/01/2026 21:39 | >>>>> | To | [email protected] | >>>>> | Cc | | >>>>> | Subject | Merge PR principles reminder | >>>>> I'm not sure if it's appropriate to bring this up, but I feel it's >>>>> necessary to raise this point. >>>>> >>>>> Recently, I've noticed that some pull requests related to Jira issues have >>>>> been merged without approval from committers and PMCs. Some even only >>>>> received AI review, without any review from other committers or PMCs. I >>>>> think this is unreasonable, and I hope this situation can be reduced in >>>>> the >>>>> future. >>>>> >>>>> Calcite is a project without a commercial company behind it, yet it has >>>>> still been able to develop healthily for many years, thanks to the efforts >>>>> of everyone in the community. This also demonstrates that Calcite's >>>>> community governance policies are sound, and we don't need to break any >>>>> fundamental principles. >>>>> >>>>> Best wishes, >>>>> Cancai >>>> >>
