PGV is unmaintained

[rmichela]

Protoc-gen-validate has no active maintainers. If you would like to take up the mantle, contact @mattklein123 here, or on #protoc-gen-validate channel in the Envoy proxy slack.

[rmichela]

#398 #74

[akshayjshah]

Hi @mattklein123 - my coworkers and I at @bufbuild would be happy to maintain PGV! It fills a huge gap in the protobuf ecosystem and deserves careful stewardship.

As a company, our incentives are well-aligned with the PGV community's: we want protobuf to be feature-complete and easy to use for personal projects, tiny startups, and giant megacorporations alike, because wider protobuf adoption increases the market for our paid schema registry product.

We'd take care not to break existing PGV users, and we can devote enough time to the project to clear out the issue backlog and possibly support additional languages. We would, of course, keep everything under an Apache 2 license. If the the current stakeholders are amenable, we can iron out a detailed transition plan.

[mattklein123]

@akshayjshah yay! Thank you! This sounds great to me. Please start with the transition plan as I doubt anyone will object but we can leave this comment thread open just in case.

[akshayjshah]

OK, can do.

Ideally, we'd transfer the repo to bufbuild/protoc-gen-validate. Github maintains redirects, so that keeps things simple: there's only one PGV repo, maintainership is clear, and anyone looking for the old URLs gets automatically redirected to the new home.

This shouldn't be too invasive for most users. Java and Python users should be unaffected, because we'd continue to publish to PyPI and Maven Central using the existing package names. I'm less familiar with the C++ code, but I think the combination of Github redirects and keeping any namespaces unchanged should be workable. Most Go users should be fine, since 99% of them are just go installing the plugin. There are some gaps in this plan (for example, we'd need to change the module name in go.mod, which would break anyone importing the code in module), but IMO they're small.

[mattklein123]

Transferring over SGTM as long as we can be reasonably sure we won't break existing deployments to the greatest extent possible.

[derekperkins]

I can't imagine a better steward than the team at buf

[derekperkins]

Is there a timeline for transferring ownership? In the meantime, could @akshayjshah be added as a maintainer to start handling some outstanding issues/prs?

[mattklein123]

There is nothing blocking this. We can do the transfer whenever. Also happy to make @akshayjshah a maintainer in the meantime. Let me know.

[akshayjshah]

The main blocker from our end is just making sure that we won't break lots of users during the transfer. @elliotmjackson has been doing some more detailed investigation to make sure that we've got our bases covered. We'll ping this issue in a bit when we're ready to transfer.

In the meantime, adding me and @elliotmjackson as maintainers sounds good!

[mattklein123]

@akshayjshah @elliotmjackson I sent you invites, then you will be maintainers. Thank you!