How to Run Tests

luci-sso uses a multi-tiered testing strategy. See Testing Architecture for how the tiers are structured and how to write new tests.

---

Unit & Integration Tests (Tiers 0–4)

These tests run inside the openwrt container using the ucode interpreter. No real router or network access is required, but the CI stack must be running:

make -C devenv up

Then run the tests:

# Run all tests
make -C devenv unit-test

# Run with detailed output
make -C devenv unit-test VERBOSE=1

# Run tests matching a pattern (regex on test name)
make -C devenv unit-test FILTER='oidc.*discovery'

# Run a specific test file or directory
make -C devenv unit-test MODULES='test/tier2/oidc_logic_test.uc'

# Select the crypto backend to test (mbedtls, wolfssl, openssl)
make -C devenv unit-test CRYPTO_LIB=wolfssl

---

End-to-End (E2E) Tests

These tests run in a Playwright-enabled Docker container and verify the full browser login flow against a Mock Identity Provider.

# Start the test stack
make -C devenv up

# Execute all browser tests
make -C devenv e2e-test

# Run tests matching a pattern
make -C devenv e2e-test FILTER='login'

# Run a specific spec file
make -C devenv e2e-test MODULES='test/e2e/01-login.spec.js'

---

Automated Watcher

You can run tests automatically whenever a file is changed in the files/, src/, or test/ directories. This requires inotify-tools installed on your host machine.

# Watch and re-run both unit and E2E tests
make -C devenv watch-tests

# Watch with filters
make -C devenv watch-tests FILTER='oidc' MODULES='test/tier2'

---

Fuzz Testing

Coverage-guided fuzzing (libFuzzer) hardens our native C components against malformed inputs.

# Run the fuzzer for the mbedtls backend
make -sC devenv fuzzer-test CRYPTO_LIB=mbedtls

See Fuzz Testing for how to analyze crashes and interpret results.