Add a log(message[, data]) method to TestContext and SuiteContext that
emits a new test:log event. Unlike test:diagnostic, which is buffered
so it is emitted in the order tests are defined, test:log is emitted
immediately, in the order tests execute, including under process
isolation where it bypasses the per-file declaration order buffer.
This gives reporters that render the test tree unbuffered a live,
attributed logging channel that captured stdout cannot provide under
concurrency.
The event carries the message, an optional opaque structured payload
that the runner passes through untouched, and the emitting test's
name, testId, parentId, nesting, and location. Built-in reporters
render it the same way they render test:diagnostic.
Signed-off-by: Moses Atlow <moshe@atlow.co.il>
PR-URL: https://github.com/nodejs/node/pull/64389
Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com>
Reviewed-By: Chemi Atlow <chemi@atlow.co.il>
Emit the standard JUnit timestamp (ISO 8601) on <testsuite> elements,
which the reporter was omitting. The suite start time is reconstructed
as end-minus-duration, because the runner reports a suite's test:start
lazily (when its first subtest reports), which would otherwise record a
time close to the suite's end.
Fixes: https://github.com/nodejs/node/issues/64028
Signed-off-by: sangwook <rewq5991@gmail.com>
PR-URL: https://github.com/nodejs/node/pull/64029
Reviewed-By: Moshe Atlow <moshe@atlow.co.il>
Reviewed-By: Jacob Smith <jacob@frende.me>
shuffleArrayWithSeed was introduced along with test order
randomization but was never called: the feature picks a random
pending subtest at dequeue time using createSeededGenerator
directly. Remove the dead helper and its export, and add direct
unit coverage for createSeededGenerator's deterministic contract.
Signed-off-by: Daijiro Wachi <daijiro.wachi@gmail.com>
PR-URL: https://github.com/nodejs/node/pull/63847
Reviewed-By: Chemi Atlow <chemi@atlow.co.il>
Reviewed-By: Jacob Smith <jacob@frende.me>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
`shouldSkipFileCoverage()` called `matchGlobPattern()` for each covered
file, and `matchGlobPattern()` builds a fresh Minimatch (a full glob
parse and regexp compile) on every call. The coverage exclude/include
globs never change during a run, so the same patterns were recompiled
once per file, which dominated the coverage report time.
Compile `coverageExcludeGlobs`/`coverageIncludeGlobs` to matchers once
per `TestCoverage` instance and reuse them for every file. Expose
`createMatcher()` from `internal/fs/glob` for this.
On a synthetic 200-test-file project this drops shouldSkipFileCoverage
from ~117ms to ~11ms (cpu-prof, isolation=none); the saving scales with
files * globs.
Refs: https://github.com/nodejs/node/issues/55103
Signed-off-by: sangwook <rewq5991@gmail.com>
PR-URL: https://github.com/nodejs/node/pull/63675
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Moshe Atlow <moshe@atlow.co.il>
Reviewed-By: Chemi Atlow <chemi@atlow.co.il>
`shouldSkipFileCoverage(url)` is invoked twice for every covered
script (once during source-map mapping and once during merge), and
the same script URL is typically reported by every worker. The
result depends only on `options.cwd`, `coverageExcludeGlobs`, and
`coverageIncludeGlobs`, all of which are fixed for the lifetime of
a TestCoverage instance, so the URL -> boolean mapping is
deterministic and safe to cache.
Add a private `#skipCache` SafeMap and split the method into a thin
caching wrapper plus a private `#computeShouldSkipFileCoverage`
that holds the original logic. Callers are unchanged.
This eliminates redundant glob and URL parsing work proportional to
the number of workers x scripts in the coverage report.
Refs: https://github.com/nodejs/node/issues/55103
Signed-off-by: sangwook <rewq5991@gmail.com>
PR-URL: https://github.com/nodejs/node/pull/63675
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Moshe Atlow <moshe@atlow.co.il>
Reviewed-By: Chemi Atlow <chemi@atlow.co.il>
Three independent bugs interacted to let a real failure-on-retry be
reported as a pass:
1. The runner's disambiguator stored the counter against the suffixed
identifier (after mutation) instead of the base key, so the counter
never advanced past 1 and every 3rd+ same-loc registration collided
on :(1).
2. The reporter had the same off-by-one when writing the state file.
3. The reporter only bumped its counter on `test:pass`, so any failing
test at a shared source location desynchronised the writer and
runner counters - on retry, the surviving failing sibling would
inherit a slot that in the previous attempt belonged to a different
(passing) sibling. Node matched by that slot, replaced `this.fn`
with a synthetic noop replay, and reported the failure as a pass.
Track the base identifier separately in the runner, bump the counter
against the base key in both the runner and the reporter, and bump the
reporter's counter on `test:fail` in addition to `test:pass`.
Fixes: https://github.com/nodejs/node/issues/63424
Signed-off-by: atlowChemi <chemi@atlow.co.il>
PR-URL: https://github.com/nodejs/node/pull/63431
Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com>
Reviewed-By: Moshe Atlow <moshe@atlow.co.il>
Add a `startCoverage` method on the `profiler` internal binding so
that V8 precise coverage can be enabled after bootstrap. The method
is idempotent against the existing bootstrap path (which creates a
V8CoverageConnection when NODE_V8_COVERAGE or
--experimental-test-coverage is set) and a no-op when the inspector
is unavailable, e.g. in the parent process of
`--test --test-isolation=process` where workers handle coverage and
Environment::should_create_inspector() returns false.
Refs: https://github.com/nodejs/node/issues/60023
Signed-off-by: sangwook <rewq5991@gmail.com>
PR-URL: https://github.com/nodejs/node/pull/63079
Reviewed-By: Chemi Atlow <chemi@atlow.co.il>
Reviewed-By: Pietro Marchini <pietro.marchini94@gmail.com>
Reviewed-By: Aviv Keller <me@aviv.sh>
Adds a `tags` option to test(), it(), suite(), and describe() that
accepts an array of string labels. Tags are canonicalized to lowercase
and inherited from suites to nested tests by union. Reporter events
expose the tag set on every test, and TestContext exposes the test's
tags via `context.tags`.
Filtering is done by literal tag name through the new
`--experimental-test-tag-filter=<tag>` flag (or `testTagFilters` on
run()). The flag may be specified more than once; tests must contain
every filter to run. Untagged tests are excluded under any positive
filter.
The tagging mechanism is gated behind a one-shot ExperimentalWarning.
Signed-off-by: atlowChemi <chemi@atlow.co.il>
PR-URL: https://github.com/nodejs/node/pull/63221
Reviewed-By: Moshe Atlow <moshe@atlow.co.il>
Reviewed-By: Pietro Marchini <pietro.marchini94@gmail.com>
Reviewed-By: Aviv Keller <me@aviv.sh>
- For imported CJS, if it's not customized by asynchronous hooks,
make sure it won't use the quirky re-invented require in all
cases.
- When the imported CJS module is customized by synchronous hooks,
in the synthetic module evalutation step, avoid calling the
respective default step again.
- Make the branching of loadCJSModuleWithModuleLoad() and
loadCJSModuleWithSpecialRequire() more explicit, and fold
the tentative fs read in the 'commonjs' translator into the
share createCJSModuleWrap() helper instead of checking it
twice in the same path.
Signed-off-by: Joyee Cheung <joyeec9h3@gmail.com>
PR-URL: https://github.com/nodejs/node/pull/62920
Fixes: https://github.com/nodejs/node/issues/63060
Reviewed-By: Paolo Insogna <paolo@cowtech.it>
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Gürgün Dayıoğlu <hey@gurgun.day>
This adds support for identifying which worker is running a test file
when tests execute concurrently, similar to JEST_WORKER_ID in Jest,
VITEST_POOL_ID in Vitest, and MOCHA_WORKER_ID in Mocha.
When running with --test-isolation=process (default), each test file
runs in a separate child process and receives a unique worker ID from
1 to N. When running with --test-isolation=none, all tests run in the
same process and the worker ID is always 1.
This enables users to allocate separate resources (databases, ports,
etc.) for each test worker to avoid conflicts during concurrent
execution.
Changes:
- Add WorkerIdPool class to manage worker ID allocation and reuse
- Set NODE_TEST_WORKER_ID environment variable for child processes
- Add context.workerId getter to TestContext class
- Add tests for worker ID functionality
- Add documentation for context.workerId
Fixes: https://github.com/nodejs/node/issues/55842
PR-URL: https://github.com/nodejs/node/pull/61394
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Pietro Marchini <pietro.marchini94@gmail.com>
Reviewed-By: Jacob Smith <jacob@frende.me>
When the test runner process is killed with SIGINT (Ctrl+C), display
which test was running at the time of interruption. This makes it
easier to identify tests that hang or take too long.
- Add `test:interrupted` event emitted when SIGINT is received
- Add `interrupted()` method to TestsStream
- Handle the event in both TAP and spec reporters
- TAP outputs: `# Interrupted while running: <test>`
- Spec outputs with yellow header and warning symbol
- Use setImmediate to allow reporter stream to flush before exit
With process isolation (default), shows the file path since the
parent runner only knows about file-level tests.
With --test-isolation=none, shows the actual test name.
PR-URL: https://github.com/nodejs/node/pull/61676
Reviewed-By: Moshe Atlow <moshe@atlow.co.il>
Reviewed-By: Chemi Atlow <chemi@atlow.co.il>
Jihwan