We're using:
- Hardhat to compile, test, and deploy our smart contracts.
- Slither, Slitherin, and Echidna, from the Trail of Bits contract security toolkit for static analysis, fuzz checking, and differential testing.
- Prettier to auto-format both Solidity and Typescript (test) code
- Solhint for Solidity linting
- ESlint for Typescript linting
These instructions assume you already have standard installations of node
, npm
, and python3
.
Set up yarn and hardhat, needed for compiling and running tests:
# If needed, install yarn
npm install -g yarn
# Clone this repo
git clone [email protected]:reserve-protocol/protocol.git
# Install packages from npm (including Solidity dependencies)
cd protocol
yarn
# Setup git hooks
yarn prepare
# Init a local .env file
cp .env.example .env
If you are going to use a Tenderly network, do the following:
- Install the tenderly cli
- Login
tenderly login --authentication-method access-key --access-key {your_access_key} --force
- Configure the
TENDERLY_RPC_URL
in your.env
file
You should also setup slither
and slitherin
. The Trail of Bits tools require solc-select. Check the installation instructions to ensure you have all prerequisites. Then:
# Install solc-select and slither
pip3 install solc-select slither-analyzer
# Include slitherin detectors within slither
pip3 install slitherin
# Install and use solc version 0.8.19
solc-select install 0.8.19
solc-select use 0.8.19
# Double-check that your slither version is at least 0.8.3!
hash -r && slither --version
# Slitherin version should be at least 0.7.0
slitherin --version
- Compile:
yarn compile
- Autoformat solidity and typescript:
yarn prettier
- Report compiled contract sizes:
yarn size
- There are many available test sets. A few of the most useful are:
- Run only fast tests:
yarn test:fast
- Run P0 tests:
yarn test:p0
- Run P1 tests:
yarn test:p1
- Run plugin tests:
yarn test:plugins
- Run integration tests:
yarn test:integration
- Run tests and report test coverage:
yarn test:coverage
- Run only fast tests:
- Lint Solidity + Typescript code:
yarn lint
- Run the Slither static checker:
yarn slither
(will include Slitherin detectors) - Run a local mainnet fork devchain:
yarn devchain
- Deploy to devchain:
yarn deploy:run --network localhost
The tests located in test/integration
will require the Mainnet Forking setup in place. This is done by setting the MAINNET_RPC_URL
variable in your local .env
. An Alchemy or Ankr node (something with archive data) is needed for Mainnet Forking to properly work. Additional information can be found here.
We use git pre-push validation to ensure that the code in our master
branch always lints, compiles, and at least passes our "fast" tests, before we even share it with each other. However, if you're working on a separate branch, and it's more practical to share code with teammates working to a different standard, you can easily comment out or otherwise modify the lines in .husky/pre-push
, and commit those along with your branch.
However, ensure that you do not change the value of .husky/pre-push
in our shared master branch; this is the appropriate set of validations for sharing code there.
We have some tooling for testing with Echidna, but it is specifically in fuzz
branch of the repo. See that branch and our echidna usage docs
See our deployment documentation.
The ToB Sliter tool is run on any pull request, and is expected to be checked by devs for any unexpected high or medium issues raised. It also includes the additional Slitherin detectors developed by Pessimistic.