capnweb-client

High-performance Rust client for Cap'n Web RPC protocol with batching and pipelining.

Overview

capnweb-client provides a feature-rich client implementation for the Cap'n Web RPC protocol. It supports automatic batching, promise pipelining, capability management, and multiple transport options.

Features

• Automatic batching: Efficiently batch multiple RPC calls
• Promise pipelining: Chain operations without waiting for intermediate results
• Multiple transports: HTTP, WebSocket, and WebTransport support
• Connection pooling: Reuse connections for better performance
• Retry logic: Automatic retry with exponential backoff
• Type-safe API: Strongly-typed client interface

Usage

Add to your Cargo.toml:

[dependencies]
capnweb-client = "0.1.0"

Basic client usage:

use capnweb_client::{Client, ClientConfig};
use serde_json::json;
use anyhow::Result;

#[tokio::main]
async fn main() -> Result<()> {
    // Create client with configuration
    let config = ClientConfig {
        url: "http://localhost:8080/rpc/batch".to_string(),
        batch_size: 10,
        batch_timeout_ms: 100,
        ..Default::default()
    };

    let client = Client::new(config)?;

    // Make RPC calls
    let result = client
        .call("calculator", "add", vec![json!(5), json!(3)])
        .await?;

    println!("Result: {}", result);

    // Batch multiple calls
    let batch = client.batch();
    let future1 = batch.call("service1", "method1", vec![]);
    let future2 = batch.call("service2", "method2", vec![]);
    batch.execute().await?;

    let result1 = future1.await?;
    let result2 = future2.await?;

    Ok(())
}

Advanced Features

Promise Pipelining

// Chain operations on promises
let promise = client.call("service", "getUser", vec![json!(123)]);
let email_promise = promise.pipeline("getEmail", vec![]);
let email = email_promise.await?;

Connection Management

// Use WebSocket for real-time communication
let ws_client = Client::websocket("ws://localhost:8080/rpc/ws").await?;

// Subscribe to events
ws_client.subscribe("events", |event| {
    println!("Received event: {:?}", event);
}).await?;

Configuration Options

• url: Server endpoint URL
• batch_size: Maximum batch size
• batch_timeout_ms: Batch window timeout
• connection_timeout_ms: Connection timeout
• retry_count: Number of retries on failure
• retry_delay_ms: Initial retry delay

License

This project is licensed under either of

• Apache License, Version 2.0, (LICENSE-APACHE or https://www.apache.org/licenses/LICENSE-2.0)
• MIT license (LICENSE-MIT or https://opensource.org/licenses/MIT)

at your option.