Update librdkafka to 2.8.0

[DoumanAsh]

No description provided.

[jmjoy]

#755

[jmjoy]

This library seems to be stagnant.

[benesch]

Indeed. There are some maintainers from Datadog who have limited but nonzero time to keep things running. Unfortunately I no longer have the time to help out myself.

[Etherdrake]

Indeed. There are some maintainers from Datadog who have limited but nonzero time to keep things running. Unfortunately I no longer have the time to help out myself.

Please merge this commit. Right now building Rust binaries is broken on Fedora 41 (Forty One).

[benesch]

@Etherdrake, while we don't have an explicit code of conduct, that kind of comment is not welcome here. There are more appropriate forums for leveling complaints against Datadog.

I can appreciate that it's frustrating to have this PR outstanding, but it's not as simple as just pressing merge. Someone with context needs to review the code and be prepared for any fallout that might result from the upgrade. In the meantime, if you need a newer version of librdkafka, you can always assume the risk and use a Git source dependency on @DoumanAsh's branch.

[Etherdrake]

@Etherdrake, while we don't have an explicit code of conduct, that kind of comment is not welcome here. There are more appropriate forums for leveling complaints against Datadog.

It was a joke obviously. But things are easily taken as an insult these days. The internet used to be a lot more jolly as I'm sure you do remember. I would also like to point out that this is an open-source project and the idea that there is some kind of censorship happening on this discussion forum over something like a cheeky reference to the infamous telephone-sales practices of a hundred billion dollar company is a bit odd.

I can appreciate that it's frustrating to have this PR outstanding, but it's not as simple as just pressing merge. Someone with context needs to review the code and be prepared for any fallout that might result from the upgrade. In the meantime, if you need a newer version of librdkafka, you can always assume the risk and use a Git source dependency on @DoumanAsh's branch.

Thanks for the suggestion!

[benesch]

If it was a joke, it was a joke at the expense of the Datadog engineers who graciously donate their time to maintain this project. It's not odd for open source projects to have codes of conduct with guidelines for behavior; see e.g. https://kubernetes.io/community/code-of-conduct/. There are other forums that are better suited for exercising your right to free speech.

Thank you for updating your comment. I hadn't actually realized that the latest version of Fedora/RHEL was shipping an OpenSSL that was incompatible with the latest release of rust-rdkafka.

[Etherdrake]

If it was a joke, it was a joke at the expense of the Datadog engineers who graciously donate their time to maintain this project.

We disagree on that. Datadog is a highly successful business with amazing logging capabilities. I never criticized their engineering.

Thank you for updating your comment. I hadn't actually realized that the latest version of Fedora/RHEL was shipping an OpenSSL that was incompatible with the latest release of rust-rdkafka.

No problem and thanks for your time and contributions to this repo.

[jmjoy]

I'm just wondering what the problem would be if upgraded to a compatible version of rdkafka?

[DoumanAsh]

rdkafka can have own bugs
I didn't do a lot of tests, but what prompted me to upgrade rdkafka is bug in admin client itself
You cannot be certain that rdkafka upgrade wouldn't affect anything else

Although I think in my tests there was no other regression, but it is not really comprehensive testing

[jdcasale]

@benesch / @fede1024 / @davidblewett

Totally understand that y'all have day jobs and you probably don't have space to make this project a top priority -- that happens.

Is there any chance we could mint a few new maintainers with merge permissions so that we can get some bandwidth dedicated toward pushing some of this stuff forward? This library is mature and very useful, so it'd be a shame if it died out because we don't have bandwidth to get stuff like dependency updates/CVE remediation prioritized.

I don't think anyone really needs/wants major architectural changes at this point, but I'd love to know that if I rely on this for something important, I'm not going to need to fork it in order to make sure I can take timely updates/bugfixes/vuln patches.

[benesch]

Hi @jdcasale—thanks for the thoughtful message! I’m afraid I don’t have the power to mint new maintainers though. Only @fede1024 does.