CLI tool meant to find flaky test by running multiple times a test harness until it fails
Both techniques requires Rust and Cargo that can be isntall following this.
Using Cargo:
bash
cargo install flaky-finder
Or by compiling it, but you will need to use the binary in target/release/flaky-finder
:
bash
git clone https://github.com/dymayday/flaky-finder.git && cd flaky-finder
cargo build --release
To run 100 times a test over 2 threads in parallel with cargo to discover a potential flakyness without stop at the first error encountered, one can run this:
bash
flaky-finder -j2 -r100 --continue "cargo test -- --nocapture release_test"
But keep in mind that this can give false results if your tests cannot be run in parallel, if they use the same network stack from the OS for example.