## Hyperlane Rust implementations ### Setup - install `rustup` - [link here](https://rustup.rs/) Note: You should be running >= version `1.69.0` of the rustc compiler, you can see that version with this command and should see similar output: ``` $ rustup --version rustup 1.26.0 (5af9b9484 2023-04-05) info: This is the version for the rustup toolchain manager, not the rustc compiler. info: The currently active `rustc` version is `rustc 1.69.0 (84c898d65 2023-04-16)` ``` ### Running Locally To run the validator, run: ```bash cargo run --release --bin validator ``` Or build and then run the binary directly: ```bash cargo build --release --bin validator ./target/release/validator ``` To run the relayer, run: ```bash cargo run --release --bin relayer ``` Or build and then run the binary directly: ```bash cargo build --release --bin relayer ./target/release/relayer ``` ### Running local binary against cloud resources (AWS KMS, S3, Postgresql etc) Building the docker image and upgrading the pod is a **slow** process. To speed up the development cycle, you can run a local binary against cloud resources. This workflow is useful for testing local changes against cloud resources. It is also useful for debugging issues in production. Example of fetching env from pod: ```bash kubectl exec fuji-hyperlane-agent-validator-0 --namespace testnet3 -c agent -- printenv > ./config/validator.fuji.env ``` Copy directory (rocks DB) from pod to local: ```bash kubectl cp testnet3/fuji-hyperlane-agent-validator-0:/usr/share/hyperlane /tmp/fuji-validator-db ``` Configure additional env variables appropriately: ```bash HYP_BASE_DB=/tmp/fuji-validator-db CONFIG_FILES=./config/testnet_config.json HYP_BASE_TRACING_FMT=pretty DATABASE_URL= # for scraper ``` Run binary with env copied from pod: ```bash env $(cat ./config/validator.fuji.env | grep -v "#" | xargs) ./target/debug/validator ``` #### Automated E2E Test To perform an automated e2e test of the agents locally, from within the `hyperlane-monorepo/rust` directory, run: ```bash cargo run --release --bin run-locally ``` This will automatically build the agents, start a local node, build and deploy the contracts, and run a relayer and validator. By default, this test will run indefinitely, but can be stopped with `ctrl-c`. ### Building Agent Docker Images There exists a docker build for the agent binaries. These docker images are used for deploying the agents in a production environment. ```bash cd rust ./build.sh ``` ### Deploy Procedure The contract addresses of each deploy can be found in `rust/config`. The agents will automatically pull in all configs in this directory. When agents are deployed to point at a new environment, they cease to point at the old ones. We **do not** continue to operate off-chain agents on old contract deploys. Contracts not supported by the agents will cease to function (i.e. messages will not be relayed between chains). Off-chain agents are **not** automatically re-deployed when new contract deploys are merged. Auto-redeploys will be implemented at some future date. ### Useful cargo commands - `cargo doc --open` - generate documentation and open it in a web browser - `cargo build` - compile the project - `cargo run --example example` - run the default executable for the current project - `cargo test` - run the tests ### Useful cargo extensions - tree - show the dependency tree. Allows searching for specific packages - install: `cargo install cargo-tree` - invoke: `cargo tree` - clippy - search the codebase for a large number of lints and bad patterns - install: `rustup component add clippy` - invoke: `cargo clippy` - expand - expand macros and procedural macros. Show the code generated by the preprocessor - useful for debugging `#[macros]` and `macros!()` - install: `cargo install cargo-expand` - invoke `cargo expand path::to::module` ### Architecture The on-chain portions of Hyperlane are written in Solidity. The rust portions are exclusively off-chain. Later, there may be on-chain rust for Near/Solana/ Polkadot. Hyperlane will be managed by a number of small off-chain programs ("agents"). Each of these will have a specific role. We want these roles to be simple, and easily described. Each of these agents will connect to a home chain and any number of replicas. They need to be configured with chain connection details and have access to a reliable node for each chain. For Ethereum and Celo connections we use [ethers-rs](https://github.com/gakonst/ethers-rs). Please see the docs [here](https://docs.rs/ethers/0.2.0/ethers/). We use the tokio async runtime environment. Please see the docs [here](https://docs.rs/tokio/1.1.0/tokio/). ### Repo layout - `hyperlane-base` - lowest dependency hyperlane utilities - contains shared utilities for building off-chain agents - this includes - trait implementations for different chains - shared configuration file formats - basic setup for an off-chain agent - `hyperlane-core` - depends on hyperlane-base - contains implementations of core primitives - this includes - traits (interfaces) for the on-chain contracts - model implementations of the contracts in rust - merkle tree implementations (for provers) - `chains/hyperlane-ethereum` - depends on hyperlane-core (and transitively hyperlane-base) - interfaces to the ethereum contracts - `chains/hyperlane-fuel` - depends on hyperlane-core - interfaces to the fuel contracts - `agents` - each of the off-chain agents implemented thus far