Hi David, Thanks for bringing this up. I understand the concern that you're bringing forward, but I think that's not the right solution here.
1. When the FLIP was brought up and discussed, this should have been a discussion topic from the beginning. If there are no committers available to help shepherd this codebase into ASF Flink, then nothing can get migrated all together. I also think it's not a good idea to say "I don't want this to be the fate of the HTTP connector". The Flink community (especially committers and PMC) is responsible for everything that's part of ASF Flink. If there are no committers willing to merge this in, then I wonder why the FLIP got accepted and if it's actually a good idea. 2. We're talking about a hypothetical scenario so far, because there's no PR yet opened for the new repository. 3. It's good that you bring this up, but ultimately the Dev mailing list can't decide on who gets committership. Only the PMC can decide that with a discussion, as listed in the bylaws [1] 4. Being a committer doesn't necessarily help with getting the releases out, because only PMC members can cast binding votes. 5. Even if I ignore everything, it sets a potentially dangerous precedent as a fast-track to committership. New committers don't get evaluated for what they're doing in the Flink community, but purely for outside contributions. What I think is the right approach, is: * There's one or more committers who are willing to help out with PR merges/reviews. * As the existing maintainers for the HTTP Connector make new changes, that work gets evaluated and considered in the PMC. * We get into the state that there's not a new Flink connector release required for every new Flink release. That was very much the intent, but given that we had changes on both the Source and Sink APIs, we had to account for more releases. I think we're currently in a better state, but that's just my quick observation and it could be that I'm wrong. Best regards, Martijn [1] https://cwiki.apache.org/confluence/display/FLINK/Flink+Bylaws On Fri, Jul 11, 2025 at 11:38 AM David Radley <david_rad...@uk.ibm.com> wrote: > Hello, > Raising this as new thread -so it is more noticeable. > > I wanted to raise a concern around the http connector. I am aware that > there are many Flink connectors and only some of them have released > versions for Flink 1.20 and even less work with Flink 2.0. I do not want > this to be the fate of the http connector. > > In order to avoid this and to have a healthy connector going forward, I > think we need the interested parties to have write access to this > repository[1]. Specifically, we should grandfather in the GetInData > committers (full disclosure I am one of these committers) as committers for > this repository. This would mean there would be motivated , interested > people who would be responsible (and accountable) to improve, fix and > release this connector. > > How might we justify this, as the people are unknown to the Flink > community: > > * We have already accepted the connector with their work in as a Flink > contribution > * We will bring in the commits in the git history so these could be > viewed as Apache contributions. > * Is there a risk in doing this? Yes, but I think there is a bigger > risk in not doing this. It seems to be a choice between a viable community > around this repo or another under resourced connector repo, with existing > Flink committers not being familiar or motivated to spend time with the > repo. So merging in PRs and creating releases will not be done in a timely > way. > > In the spirit of community over code, I think we should bring in the > GetInData committers as Flink committers, enhancing the community to ensure > that this is a healthy connector going forward and it gets the love it > deserves! > > Other connectors that are not getting Flink 1.20 and v2 releases will also > need a group of committers to ensure that they get released; if there is > not an identifiable group of people to keep these connectors healthy, I > suggest we look to mothball (remove the docs from Flink master) these > connectors until such time that there are committers that can support them. > > Also, I like the idea[2] Danny raised. We can try this with the HTTP > connector. > > Thoughts? > > Kind regards, David. > [1] https://github.com/apache/flink-connector-http > [2] https://lists.apache.org/thread/byy8fgkr8dbrrdv2nxpxkrz18h904r9b > > > Unless otherwise stated above: > > IBM United Kingdom Limited > Registered in England and Wales with number 741598 > Registered office: Building C, IBM Hursley Office, Hursley Park Road, > Winchester, Hampshire SO21 2JN >
