implement protoc-gen-prost-validate

[Adphi]

Implementation of protoc-gen-prost-validate using prost-validate.

closes #106

[Adphi]

Hi @neoeinstein !

Could you please take a look at this PR when you have a chance ?
Let me know if you have any feedback or suggestions.

Thanks in advance!

[fenollp]

FYI @neoeinstein we're a few people already using this and it works great :)

[siennathesane]

Hey hey, I'm a new maintainer now. I've not used prost-validate. I'm open to this on principal but I want to do some research so I can give you a good review.

[siennathesane]

I'm on the fence about this. It feels like the maintenance burden will be high, esp since there's an 8.5k line test case file with no documentation that I can't tell is hand-generated or not. This feels risky from a maintenance and stability standpoint.

@neoeinstein what do you think?

[Adphi]

@siennathesane, thanks for taking a look, that’s a fair concern.

For context, the large test file isn’t unique to this PR. The tests are taken directly from prost-validate-tests which themselves are adapted from the original protoc-gen-validate harness tests.

So this is a long-standing, well-tested corpus reused across multiple implementations, rather than new surface area introduced here. The goal was to reduce risk by aligning with existing behavior.

We’re also happy to take on maintenance ourselves, either within this repo or by maintaining our own fork. The main blocker to doing the latter is needing the protoc-gen-prost-validate crate name to publish it.

[siennathesane]

Hmm... I think there's value for the community, but my biggest concern is the maintenance burden. I would rather own the maintenance of it than have forks so that we can keep stuff useful for everyone and keep all the crates in one place.

I'll make you a deal. If you can come up with a MAINTAINERS.md (or some appropriate name) document for this crate, with all of the instructions on how to maintain the test files, the protobufs, and other general maintenance tasks, I'll approve it. It doesn't need to be 100% perfect, but enough that someone like me can look at it, understand what to do, and get it done.

I'm willing to meet in the middle on this because I think there's value for the community, and I want this project to continue to be the source of that value in the Buf ecosystem. But I need a way to make sure that I, and others, are able to maintain it across versions without anyone having to be an expert, or to put you on the hook for maintaining it. To me, documenting the process feels like a good middle ground.

There are merge conflicts, CI will need to pass (don't worry about the vet job), and other stuff to get it merged, but that's minor imo.

What do you think?