https://pulsar.apache.org/ logo
Join Slack
Powered by
Hello <@ULBJ01G6N> and <@U680ZCXA5>. Appreciate if...
# dev
p
Hello @Lari Hotari and @merlimat. Appreciate if you can explain me if this behavior is expected in Pulsar community: github.com/apache/pulsar/issues/25978#… Thanks
l
It's not. I'm sorry that it happened. I think that Asthenia0412 is a bot and violating the ASF generative tooling guidance of "Fully autonomous agents (e.g. OpenClaw-style bots) opening PRs or filing reports on their own are not acceptable — the AI assists, the human is accountable." I've reminded the account in these comments: github.com/apache/pulsar/pull/26729#… github.com/apache/pulsar/pull/26932#… If there's no response, it's possible that we'd need to block the account although they must be spending tokens for a good purpose in any case. Some of the PRs generated with bots could be useful, but a principle is a principle. If a human doesn't show up, I'll push to block the user.
For fully automated PRs, it's better to just use a coding agent yourself than training someone else's bot.
In general humans are present in ASF with their own name and identity. I personally don't like that some users hide behind a handle and don't reveal their name and identity. Such users will never gain committer status regardless of useful commits. In ASF, releases should only be performed by persons that have established a personal relationship with a PMC member. There should be a true identity check of a ID card before adding the user's signing key to the KEYS file. However, this principle is probably not followed in many ASF projects. I once got my identity checked as a ASF conference, but the key signing got lost for some reason and the web-of-trust type of verification never got established. It's probably a similar story for many others.
contributions to 192 repositories in the last month or so
@Pavel Zeger Another reason why this happend and your PR was missed was that github.com/apache/pulsar/issues/25978 didn't have a link to your issue. In the PR description, you should reference the issue with "Fixes #xxxxx" or "Closes #xxxxx" so that it gets linked.
I added a short note in github.com/apache/pulsar/issues/25978#… . Adding the link to a PR should prevent this from happening again although I think that the PR is mostly coming from a user that uses a bot without human intervention to create PRs.
@Pavel Zeger A mistake on my side was that I didn't assign this issue to you since you had added a check box to "I'm willing to submit a PR!". I'm sorry about that.
more similar behavior from the same bot/user: https://github.com/apache/pulsar/pull/26934
m
I think it’s just going through open issues and creating PRs for those..
l
yes, an automated bot without human oversight.
m
yes, looks like that
l
they must have a lot of tokens to spend. PRs to 192 repositories recently.
m
probably some testing on models…
though, I don’t see way to block accounts on the repo. we prob need to ask ASF infra
l
... or training models
when I have time, I'll ask if someone in ASF has come up with a good solution for this. I'm pretty sure Jarek in Airflow has something in place.
having some sort of "human test" or captcha could be useful in categorizing PRs. I don't mind if someone likes to burn tokens in creating PRs. 😀
it wouldn't be too hard to guess from the GH profile if the user is a bot.
s
I think even robots should check before they work. • Pull the latest master branch, keep only the test code to see if the problem still exists. • Is there already a corresponding issue, and has the issue author checked “I’m willing to submit a PR!“. • Check if there is already a PR that is handling the relevant issue. However, I think we can also “defend” at the same time, check at the beginning of the review whether there is a relevant issue or a PR that solves the same problem. Such tasks used to be quite laborious before AI; I think now they can still be checked very quickly. Now AI is developing too fast, and it’s easy to generate various problems.😂
👍 1
l
@SiNan Liu good points. could you update CONTRIBUTING.md to contain these details?
👀 1
s
github.com/apache/pulsar/pull/26964 @Lari Hotari The CONTRIBUTING.md has been updated. Is there anything else to add or modify?
l
@SiNan Liu thanks, LGTM. we can wait a bit to allow @merlimat to check it too.
s
Yes, let‘s see what else you have to add. I remember @Yunze Xu wrote an article some time ago. Let's see if there is anything to add.
p
@Lari Hotari Thank you for the explanation and the apology, and for being willing to look at how the community rules are applied. I appreciate it. I'm glad raising this led to an investigation that turned up something useful. I do still have doubts about whether the community can keep up with the current pace of development. Several of my PRs and PIPs weren't discussed or voted on, so for now I'd rather not start anything new while those are still open. I'd like to understand whether this review bottleneck can be addressed, and how I could help move things forward, whether that's reviewing other contributors' PRs, helping with triage, or something else that would be useful.
l
There's currently an effort lead by @merlimat to cut the backlog of PRs and PIPs, thread https://lists.apache.org/thread/cgy7w0qgnq8s7j0ykyzgrn3hj4nb67wf
We are all volunteering in this project and everyone can impact the project direction regardless of committer status. "The Apache Way" is a nice introduction: https://www.apache.org/theapacheway/

https://www.youtube.com/watch?v=Dx5kQnVFo7E▾

describes what we all hope it would be so that everyone is welcome to contribute and make an impact.
Open in Slack
PreviousNext
Powered by