We use cargo-llvm-cov to generate coverage reports for tests.
cargo +stable install cargo-llvm-cov --locked# Set the environment variables needed to get coverage.
# This command sets the RUSTFLAGS and other environment variables
source <(cargo llvm-cov show-env --export-prefix)It should note that in certain cases, you might want to add additional flags to the RUSTFLAGS environment variable. For example, on certain version of OSX, you might need to add the following flag:
echo $RUSTFLAGS
# You will see something like this: `-C instrument-coverage --cfg=coverage --cfg=trybuild_no_target`
export RUSTFLAGS="${RUSTFLAGS} -L /opt/homebrew/opt/bzip2/lib -l bz2"# Remove artifacts that may affect the coverage results.
# This command should be called after show-env.
cargo llvm-cov clean --workspace
# Above two commands should be called before build binaries.
cargo build # Build rust binaries, binaries would be in target/debug/*Run unit tests:
cargo nextest run --no-fail-fast --workspace
cargo nextest run --no-fail-fast -p cfx-addr --no-default-featuresRun integration tests:
It should be noted that we compile the binary in debug mode, so the performance is not good. You might need to change parallel parameters if frequent io error or timeout error occurs.
# Run integration tests
# Change -n to control the number of tests running in parallel.
pytest integration_tests/tests -vv -n 6 --dist loadscope --conflux-binary $(pwd)/target/debug/conflux
# Set up benchmark binary path before running `python tests/test_all.py`
export CONFLUX_BENCH=$(pwd)/tools/consensus_bench/target/debug/consensus_bench
# Run additional tests.
# Use --max-workers and --max-nodes to control the number of workers and nodes.
python tests/test_all.py --conflux-binary $(pwd)/target/debug/conflux*.profraw files will be generated in ./target/
cargo llvm-cov report --html --failure-mode=all # Generated report will be in `./target/llvm-cov/html/index.html`