Development

Development

This page covers compiling, testing, and benchmarking the etr codebase.

Prerequisites

To compile and run all tests locally, you need:

• The Rust toolchain (rustup.rs)
• just (command runner)
• tmux (required for local end-to-end tests)
• OpenSSH client/server with passwordless login to localhost configured (required for E2E tests)

Building

# Debug build
cargo build

# Release build
cargo build --release

Running Tests

We maintain a comprehensive suite of static checks, unit tests, and automated end-to-end tests:

# Run formatting and clippy lints
just check

# Run unit and integration tests (78 tests)
just test

# Run local end-to-end session and reconnect tests (launches etr in tmux)
just e2e-local

# Run local end-to-end tests for -L port forwarding (TCP + UDP, IPv4 + IPv6, reconnect)
just e2e-forward-local

# Run local end-to-end tests for -R port forwarding (TCP + UDP, IPv4 + IPv6, reconnect)
just e2e-reverse-local

# Run local end-to-end test for --env variable forwarding
just e2e-env-local

# Regression test: concurrent UDP senders through -L (v0.4.9 per-sender routing)
just e2e-udp-concurrent

# Run throughput/memory stress tests
just stress-local

Running Benchmarks

We use criterion to benchmark core performance-sensitive areas of the codebase, including TLS certificate generation, connection handshakes, round-trip latency, and throughput.

# Run the benchmark suite
just bench