# Testing signals — nodejs/node

> Canonical page: https://codevetter.com/unpack/node/testing-signals

# Testing signals — nodejs/node

Scanned commit 976a36a5b6ce737c9e695f6cde8e141427f13737 (2026-09-22). Generated by codevetter codevetter 1.12.0 on 2026-09-22.

Testing is dominated by the bespoke Python harness `tools/test.py`, which discovers suites under `test/<suite>/`, applies per-suite `.status` files for SKIP/SLOW/FLAKY expectations, and shells out to the built `node` binary per test file. There is no Vitest/Jest; the stack tag is misleading. Native code is covered by `cctest` (googletest) and the `js-native-api`/`node-api`/`addons` suites, which require a build step before they can run. Coverage is collected by nyc/c8 with `.nycrc` thresholds of 95% lines/statements and 93% branches, uploaded to Codecov from two Linux workflows. The new subsystems are well covered at the JS level — ~121 `test-vfs-*` files in `test/parallel` alone plus a dedicated `test/ffi/` suite — but SEA-specific paths are explicitly excluded from coverage.
- The harness is `tools/test.py`, a Python 3 runner invoked as `./tools/test.py --flaky-tests keep_retrying -p actions -j 4`; suites are directories under `test/` and each may carry a `<section>.status` file read by `ReadConfigurationInto`. — [tools/test.py#L860-870](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/tools/test.py#L860-L870) [tools/test.py#L1391](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/tools/test.py#L1391) [.github/workflows/test-linux.yml#L93-99](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/.github/workflows/test-linux.yml#L93-L99)
- `test/root.status` encodes mode-dependent expectations (e.g. a long list of `parallel/test-*: SLOW` entries under `[$mode==debug]`), and `test/ffi/ffi.status` skips `test-ffi-calls`, `test-ffi-shared-buffer`, and `test-ffi-weakref-calls` on Solaris/SmartOS where bundled libffi callbacks crash. — [test/root.status#L1-30](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/root.status#L1-L30) [test/ffi/ffi.status](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/ffi/ffi.status)
- `test/parallel` is by far the largest suite (~4976 files), followed by `test/fixtures` (~6108 support files), `test/es-module` (246), `test/addons` (185), `test/sequential` (125), `test/module-hooks` (114), and `test/test-runner` (113). — [test/parallel](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/parallel) [test/es-module](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/es-module) [test/module-hooks](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/module-hooks)
- VFS has roughly 121 dedicated files in `test/parallel` covering per-API parity (`test-vfs-fs-mkdirSync`, `test-vfs-fs-copyFileSync`, `test-vfs-fs-createReadStream`), edge cases (`test-vfs-bigint-position`, `test-vfs-ctime-update`), and an explicit `test-vfs-fs-matches-real-fs` differential test. — [test/parallel/test-vfs-fs-matches-real-fs.js](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/parallel/test-vfs-fs-matches-real-fs.js) [test/parallel/test-vfs-bigint-position.js](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/parallel/test-vfs-bigint-position.js) [test/parallel/test-vfs-fd.js](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/parallel/test-vfs-fd.js)
- FFI has its own compiled suite `test/ffi/` with a `fixture_library`, a `testcfg.py`, and permission-focused tests: `test-ffi-permissions.js`, `test-ffi-permission-allow-cli.js`, `test-ffi-permission-audit.js`, `test-ffi-permission-drop.js`, plus `test-ffi-vfs.js` for the load-from-bytes path. — [test/ffi/test-ffi-permissions.js](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/ffi/test-ffi-permissions.js) [test/ffi/test-ffi-permission-audit.js](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/ffi/test-ffi-permission-audit.js) [test/ffi/test-ffi-vfs.js](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/ffi/test-ffi-vfs.js) [test/ffi/testcfg.py](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/ffi/testcfg.py)
- Running the JS suites requires native build steps first: `make jstest` depends on `build-addons build-js-native-api-tests build-node-api-tests build-sqlite-tests build-ffi-tests` — running `tools/test.py` on a clean tree will fail those suites. — [Makefile#L323](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/Makefile#L323) [Makefile#L343-361](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/Makefile#L343-L361)
- `make cctest` runs the C++ googletest binary built from `test/cctest/` (37 files) and is a separate gate from the JS suites. — [Makefile#L304](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/Makefile#L304) [test/cctest](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/cctest) [deps/googletest](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/deps/googletest)
- CI suite composition is `CI_JS_SUITES = $(JS_SUITES) pummel` and `CI_NATIVE_SUITES = $(NATIVE_SUITES) benchmark`, with `doctool` conditionally appended; `make test-ci` runs all three plus the prerequisite builds. — [Makefile#L610-615](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/Makefile#L610-L615) [Makefile#L644-650](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/Makefile#L644-L650)
- `.github/workflows/test-linux.yml` builds on ubuntu-24.04 and ubuntu-24.04-arm with clang-19, Python 3.14, Rust 1.88 and sccache, runs `make build-ci` then `make test-ci -j1 TEST_CI_ARGS="-p actions --measure-flakiness 9"`. — [.github/workflows/test-linux.yml#L40-92](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/.github/workflows/test-linux.yml#L40-L92)
- CI asserts test hygiene two ways: a `git add -A && git diff --exit-code --staged` step fails the build if running tests dirtied the tree, and the whole suite is re-run from a directory named with spaces, quotes, and non-ASCII characters to catch path-handling bugs. — [.github/workflows/test-linux.yml#L89-99](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/.github/workflows/test-linux.yml#L89-L99)
- Coverage thresholds are declared in `.nycrc` — 95% lines, 95% statements, 93% branches — over `lib/` only, with `test/`, `tools/`, `benchmark/`, and `deps/` excluded. — [.nycrc](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/.nycrc)
- Codecov is configured to wait for two coverage builds (`coverage-linux.yml` and `coverage-linux-without-intl.yml`; the Windows one is noted as manually disabled) and both project and patch status gates are turned off, so coverage is informational rather than blocking. — [codecov.yml](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/codecov.yml) [.github/workflows/coverage-linux.yml](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/.github/workflows/coverage-linux.yml) [.github/workflows/coverage-linux-without-intl.yml](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/.github/workflows/coverage-linux-without-intl.yml)
- `test/common/index.js` exposes capability predicates that tests use to self-skip — `hasCrypto`, `hasIntl`, `hasFullICU`, `hasTemporal`, `hasDtls`, `hasQuic`, `hasSQLite`, `hasFFI`, `hasInspector`, `hasLocalStorage` — so conditionally-compiled bindings do not produce false failures. — [test/common/index.js#L1008-1033](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/common/index.js#L1008-L1033)
- SEA-plus-VFS initialisation is explicitly uncovered: `internal/vfs/sea.js` opens with `/* c8 ignore start -- SEA VFS initialization requires an actual SEA binary */`, so that path is exercised only by the separate `test/sea` suite (43 files), not by line coverage. — [lib/internal/vfs/sea.js#L22](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/lib/internal/vfs/sea.js#L22) [test/sea](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/test/sea)
- Recent commit history shows active flake management (`test: deflake test-bench-errors`, `test-http2-debug`, `test-benchmark-crypto`) alongside a `label-flaky-test-issue.yml` workflow and `--measure-flakiness 9` in CI. — [.github/workflows/label-flaky-test-issue.yml](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/.github/workflows/label-flaky-test-issue.yml) [.github/workflows/stress-test.yml](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/.github/workflows/stress-test.yml) [.github/workflows/test-linux.yml#L92](https://github.com/nodejs/node/blob/976a36a5b6ce737c9e695f6cde8e141427f13737/.github/workflows/test-linux.yml#L92)

Every claim links to its cited files at the pinned commit. Citations identify source evidence; they do not establish runtime correctness. Independent CodeVetter scan — not reviewed by upstream maintainers.

## Public product links

- [CodeVetter](https://codevetter.com/)
- [Download](https://codevetter.com/download)
- [Documentation](https://codevetter.com/docs/)
- [Source](https://github.com/Codevetter/codevetter)