[question] message validation

Hey team Buf, we are loving what you are building, thank you sooo much!!!

It was a breeze switching from @improbable-eng/grpc-web to connect-web. We are looking to transitioning to connect-go in the future (still using @improbable-eng’s grpcweb).

The question here is regarding message validation. We are wondering if you are at all considering implementing message validation as a part of the connect ecosystem?

(Just for the background, we are early in the development process, and are currently evaluating envoyproxy/protoc-gen-validate for the backend and looking into colinhacks/zod with fabien0102/ts-to-zod for our frontend.)

[fubhy]

I came here looking for the same thing (typescript-first runtime message validation with zod or similar).

I’ll have a bit of time by the end of this week and meant to play around with connect-web and protobuf-es (currently, we are using protobuf-ts and have a bespoke zod based validation solution in place there). I’ll give protoplugin a go then and see how far I can get. Will share any relevant progress here in case that’s useful for others.

FYI: I’m thinking of generating the code with PGV in a way that further refines the generated types (if possible, haven’t looked into it yet) according to the validation schema.

[fubhy]

@johynpapin I pushed a first “work in progress” to that repository. Things are likely going to change quite a bit as I continue to experiment so be aware of that if you want to play around with it already.

I also haven’t implemented any runtime bindings yet. Not sure how I want to integrate it with connect-web (client side validation) and connect-node (server side validation) yet (interceptors?).

I might add another generator that uses grpc service definitions to generate a schema registry and generic validation function (that can validate any message).

For now you have to do that manually (MyMessageSchema.parse(message)).

I am not particularly fond of the string based code generation and how I used it here … I think I would rather like to print some kind of abstract, composed AST. Particularly with something like the zod API that is designed for chaining and composing function calls that would probably make the generator code quite a bit more digestable.

Anyways, it’s a start …

/cc @timostamm