Thanks to a tip from an Apache Arrow committer, I found another way to check if a GH user was a committer. This solves the issue I was facing for the "triage" label workflow.
For future automations, we may still consider including a list of committers in the source (whether on an orphaned branch or trunk). There is actually precedent for this in Apache Airflow https://github.com/apache/airflow/blob/main/dev/breeze/src/airflow_breeze/global_constants.py#L437 Anyways, thanks for the pointers everyone :) -David A On Wed, Dec 4, 2024 at 1:22 PM David Arthur <mum...@gmail.com> wrote: > Thanks for the pointers, everyone. Lots of TIL :) > > The kafka-committers team is a "secret" team in GH, so I don't think the > GH automation will be able to access it. I'll follow up with Infra to see > if there's a way. That would definitely be the simplest solution. > > Mickael, the idea behind the orphaned branch compared to trunk is that we > wouldn't have to worry about it being synchronized to branches. Git-config > has a "--glob" option which lets you specify a ref and file name. For > example, "git config --glob asf-committers:commiters.txt --get > committers.mumrah.name" would work from any branch (as long as the > asf-committers ref was pulled locally). > > -David > > On Wed, Dec 4, 2024 at 1:02 PM Divij Vaidya <divijvaidy...@gmail.com> > wrote: > >> The committer information is also publicly available in a json format at >> https://whimsy.apache.org/public/public_ldap_projects.json >> >> We can choose to parse the json to extract the committers. It won't >> require >> any authentication to generate our file. >> >> -- >> Divij Vaidya >> >> >> >> On Wed, Dec 4, 2024 at 10:56 AM Mickael Maison <mickael.mai...@gmail.com> >> wrote: >> >> > Hi David, >> > >> > Yes there's a GitHub team: >> > https://github.com/orgs/apache/teams/kafka-committers >> > >> > In case we still need to store the information in the repo, why do you >> > want to put it in a separate branch and not directly in trunk? >> > >> > Thanks, >> > Mickael >> > >> > On Wed, Dec 4, 2024 at 5:47 AM David Jacot <david.ja...@gmail.com> >> wrote: >> > > >> > > Hi David, >> > > >> > > I think that there is a « kafka committer » team within the Apache >> > > organization. You may be able to query it. It seems to be >> automatically >> > > populated based on the roaster data. >> > > >> > > Best, >> > > David >> > > >> > > Le mar. 3 déc. 2024 à 23:52, David Arthur <mum...@gmail.com> a écrit >> : >> > > >> > > > Chia-Ping, >> > > > >> > > > > Do we have documentation for the manual steps? If so, I'm + 1 to >> > create >> > > > an >> > > > orphaned branch to store the extra committer list, and could you >> > please add >> > > > this new step to the document. >> > > > >> > > > I don't see anything in the wiki, but I'd be happy to write that up. >> > > > >> > > > > Or you are talking about "approvers" >> > > > >> > > > Ah, yea I was thinking of "Approvers: " (which is a future >> discussion). >> > > > Although, I suppose the same technique could be applied to any >> > contributor >> > > > (committer or not) as a way to designate their preferred name + >> email >> > in >> > > > the commit message. >> > > > >> > > > --- >> > > > >> > > > Divij, doh! Thanks. I was looking in the old "phonebook" application >> > and >> > > > not Whismy. Since the API requires authentication, it might be >> > difficult to >> > > > integrate into our automations. Maybe we could use the API to >> extract >> > the >> > > > necessary data for the "committers" file. It also looks like the >> > > > "githubUsername" is sparsely populated, so maybe some manual labor >> is >> > > > involved regardless. >> > > > >> > > > >> > > > -David A >> > > > >> > > > On Tue, Dec 3, 2024 at 4:36 PM Divij Vaidya < >> divijvaidy...@gmail.com> >> > > > wrote: >> > > > >> > > > > Hey David >> > > > > >> > > > > You can find the authoritative list of committers and PMC members >> > here - >> > > > > https://whimsy.apache.org/roster/committee/kafka >> > > > > If you click on usernames here, you can find the corresponding >> > GitHub ID >> > > > as >> > > > > well. >> > > > > >> > > > > Whimsy has its API documented at >> https://whimsy.apache.org/docs/api/ >> > , >> > > > > hence, it may be possible to query the API and get the information >> > > > directly >> > > > > from the source of truth. I haven't tried this approach but >> throwing >> > out >> > > > an >> > > > > option out there. >> > > > > >> > > > > -- >> > > > > Divij Vaidya >> > > > > >> > > > > >> > > > > >> > > > > On Tue, Dec 3, 2024 at 9:51 PM Chia-Ping Tsai <chia7...@gmail.com >> > >> > > > wrote: >> > > > > >> > > > > > hi David, >> > > > > > >> > > > > > (I copy your response from PR to here) >> > > > > > >> > > > > > > There's already a lot of manual steps involved in becoming a >> > > > committer. >> > > > > > >> > > > > > Do we have documentation for the manual steps? If so, I'm + 1 to >> > create >> > > > > an >> > > > > > orphaned branch to store the extra committer list, and could you >> > please >> > > > > add >> > > > > > this new step to the document. >> > > > > > >> > > > > > > Generating "Reviewers:" strings automatically. >> > > > > > >> > > > > > We already have `reviewers.py` which can generate the strings. >> Or >> > you >> > > > are >> > > > > > taking about "approvers"? ( >> > > > > > >> https://github.com/apache/kafka/pull/17881#discussion_r1854582483) >> > > > > > Best, >> > > > > > Chia-Ping >> > > > > > >> > > > > > David Arthur <mum...@gmail.com> 於 2024年12月4日 週三 上午4:40寫道： >> > > > > > >> > > > > > > Hey folks, as part of >> https://github.com/apache/kafka/pull/17881 >> > I >> > > > > found >> > > > > > > myself in need of a list of github usernames that belonged to >> > Kafka >> > > > > > > committers. As far as I know, there is no such mapping defined >> > > > > anywhere. >> > > > > > > >> > > > > > > I would like to create a new branch that contains a config >> file >> > that >> > > > > lets >> > > > > > > us map GH users to their display name and emails. I have >> already >> > > > > compiled >> > > > > > > the data on my fork >> > > > > https://github.com/mumrah/kafka/tree/asf-committers. >> > > > > > > This was derived from the ASF roster and some manual searching >> > in the >> > > > > Git >> > > > > > > log and GitHub PRs. >> > > > > > > >> > > > > > > Chia-Ping raised a good point about divergence between this >> data >> > and >> > > > > the >> > > > > > > committers.html file >> > > > > > > >> > https://github.com/apache/kafka/pull/17881#issuecomment-2513252954. >> > > > > > > >> > > > > > > In addition to the PR triage workflow, I can think of some >> > additional >> > > > > use >> > > > > > > cases: >> > > > > > > >> > > > > > > * Generating committers.html from this data set >> > > > > > > * Generating "Reviewers:" strings automatically. >> > > > > > > >> > > > > > > I was wondering what folks thought about this. I can't easily >> > raise a >> > > > > PR >> > > > > > > for review since the proposal here is to create a new orphaned >> > > > branch. >> > > > > > > >> > > > > > > Thanks! >> > > > > > > David A >> > > > > > > >> > > > > > >> > > > > >> > > > >> > > > >> > > > -- >> > > > David Arthur >> > > > >> > >> > > > -- > David Arthur > -- David Arthur
