feat(proto): generate generic GRPC API implementations

[kkovaacs]

This PR adds the build-time generation of generic GRPC API implementations for all services discussed in PR #1742.

For each method we're generating a trait like this:

#[tonic::async_trait]
pub trait Status {
    type Input;
    type Output;

    fn decode(request: ()) -> tonic::Result<Self::Input>;

    fn encode(output: Self::Output) -> tonic::Result<crate::generated::validator::ValidatorStatus>;

    async fn handle(&self, input: Self::Input) -> tonic::Result<Self::Output>;

    async fn full(
        &self,
        request: (),
    ) -> tonic::Result<crate::generated::validator::ValidatorStatus> {
        let input = Self::decode(request)?;
        let output = self.handle(input).await?;
        Self::encode(output)
    }
}

The idea is that the actual implementation should consist of implementing all per-method traits for the API server (most of the time just decode / encode / handle), and then the provided generic implementation for the API can be used:

pub trait ApiService: Status + SubmitProvenTransaction + SignBlock {}

impl<T> ApiService for T
where
    T: Status,
    T: SubmitProvenTransaction,
    T: SignBlock,
{
}

#[tonic::async_trait]
impl<T> crate::generated::validator::api_server::Api for T
where
    T: ApiService,
    T: Send,
    T: Sync,
    T: 'static,
{
    async fn status(
        &self,
        request: tonic::Request<()>,
    ) -> tonic::Result<tonic::Response<crate::generated::validator::ValidatorStatus>> {
        #[allow(clippy::unit_arg)]
        <T as Status>::full(self, request.into_inner()).await.map(tonic::Response::new)
    }

    async fn submit_proven_transaction(
        &self,
        request: tonic::Request<crate::generated::transaction::ProvenTransaction>,
    ) -> tonic::Result<tonic::Response<()>> {
        #[allow(clippy::unit_arg)]
        <T as SubmitProvenTransaction>::full(self, request.into_inner())
            .await
            .map(tonic::Response::new)
    }

    async fn sign_block(
        &self,
        request: tonic::Request<crate::generated::blockchain::ProposedBlock>,
    ) -> tonic::Result<tonic::Response<crate::generated::blockchain::BlockSignature>> {
        #[allow(clippy::unit_arg)]
        <T as SignBlock>::full(self, request.into_inner())
            .await
            .map(tonic::Response::new)
    }
}

In this form this PR should be a no-op, because the none of our API implementations have been moved to this new model. That will be done in multiple follow-up PRs.

[Copilot]

Pull request overview

This PR extends the crates/proto build-time code generation to also emit generic, per-RPC-method server-side traits and blanket tonic server implementations, aiming to standardize gRPC server implementations across all services referenced in PR #1742.

Changes:

• Update crates/proto/build.rs to generate additional server facade modules from protobuf descriptors (plus regenerate mod.rs files).
• Add build-time dependencies needed for code generation and descriptor introspection (codegen, prost-types).
• Run rustfmt over generated Rust sources at build time.

Reviewed changes

Copilot reviewed 3 out of 4 changed files in this pull request and generated 4 comments.

File	Description
crates/proto/build.rs	Adds server module generation from descriptors, new mod.rs generation logic, and build-time rustfmt pass.
crates/proto/Cargo.toml	Adds build dependencies (codegen, prost-types) needed by build.rs.
Cargo.toml	Pins prost-types to align with pinned prost versions in the workspace.
Cargo.lock	Locks new dependency resolutions for codegen and prost-types.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 4 changed files in this pull request and generated 3 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[kkovaacs]

I've added code generation for server streaming. The traits we're generating for these look like this (using the mempool subscription as an example):

#[tonic::async_trait]
pub trait MempoolSubscription {
    type Input;
    type Item;
    type ItemStream: tonic::codegen::tokio_stream::Stream<Item = tonic::Result<Self::Item>>
        + Send
        + 'static;

    fn decode(request: ()) -> tonic::Result<Self::Input>;

    fn encode(item: Self::Item) -> tonic::Result<crate::generated::block_producer::MempoolEvent>;

    async fn handle(&self, input: Self::Input) -> tonic::Result<Self::ItemStream>;

    async fn full(
        &self,
        request: (),
    ) -> tonic::Result<
        std::pin::Pin<
            Box<
                dyn tonic::codegen::tokio_stream::Stream<
                        Item = tonic::Result<crate::generated::block_producer::MempoolEvent>,
                    > + Send
                    + 'static,
            >,
        >,
    > {
        use tonic::codegen::tokio_stream::StreamExt as _;
        let input = Self::decode(request)?;
        let stream = self.handle(input).await?;
        Ok(Box::pin(stream.map(|item| item.and_then(|i| Self::encode(i)))))
    }
}

And then the implementation can be something like this:

#[tonic::async_trait]
impl proto::server::block_producer_api::MempoolSubscription for BlockProducerRpcServer {
    type Input = ();
    type Item = MempoolEvent;
    type ItemStream = MempoolEventSubscriptionStream;

    fn decode(_request: ()) -> Result<Self::Input, tonic::Status> {
        Ok(())
    }

    fn encode(output: Self::Item) -> tonic::Result<proto::block_producer::MempoolEvent> {
        Ok(proto::block_producer::MempoolEvent::from(output))
    }

    async fn handle(&self, _request: Self::Input) -> tonic::Result<Self::ItemStream> {
        let subscription = self.mempool.lock().await.lock().await.subscribe();
        let subscription = ReceiverStream::new(subscription);

        Ok(MempoolEventSubscriptionStream { inner: subscription })
    }
}

struct MempoolEventSubscriptionStream {
    inner: ReceiverStream<MempoolEvent>,
}

impl tokio_stream::Stream for MempoolEventSubscriptionStream {
    type Item = tonic::Result<MempoolEvent, tonic::Status>;

    fn poll_next(
        mut self: std::pin::Pin<&mut Self>,
        cx: &mut std::task::Context<'_>,
    ) -> std::task::Poll<Option<Self::Item>> {
        self.inner.poll_next_unpin(cx).map(|x| x.map(tonic::Result::Ok))
    }
}

[Copilot]

Pull request overview

Copilot reviewed 3 out of 4 changed files in this pull request and generated 1 comment.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The ServerStream::as_trait generator emits a full body line with unbalanced parentheses: Ok(Box::pin(stream.map(...))) is missing one closing ). As written, the generated code will not compile for server-streaming RPCs (e.g. MempoolSubscription). Add the missing closing parenthesis in the string passed to .line(...).

Suggested change

	.line("Ok(Box::pin(stream.map(|item| item.and_then(|i| Self::encode(i)))))");
	.line("Ok(Box::pin(stream.map(|item| item.and_then(|i| Self::encode(i))))))");