test_runner: attribute global console output to active tests

[hanityx]

When tests run under process isolation, output from the built-in global console reaches the parent through the child process's stdout and stderr streams. By the time the parent emits test:stdout or test:stderr, it no longer knows which test produced the write. Custom reporters therefore cannot associate the output with a test, and it can arrive before the matching test:dequeue. Depending on how the pipe is drained, writes from different tests can also be combined into a single raw event.

The test runner installs an internal write hook on the built-in global console in child processes that use the V8 serializer. While a non-root test or suite is active, the formatted message is emitted through its TestsStream with testId, parentId, name, nesting, and source location metadata. These events bypass the per-file declaration-order buffer, so they stay in execution order after the matching test:dequeue.

This follows the approach suggested by cjihrig in #53103 (comment).

context.log() and context.diagnostic() remain the explicit reporting APIs. The hook also covers global console calls made by code under test or third-party modules, without requiring those callers to use the test context APIs. The existing test:stdout and test:stderr event types are preserved.

Scope

• Only test runner child processes that use the V8 serializer install the hook. Ordinary Node.js processes do not.
• Output that the hook does not attribute keeps its existing file-scoped path: direct process.stdout and process.stderr writes, native output, separately constructed Console instances, top-level and root-hook output, output written while suites are being defined, and asynchronous output after a test has finished.
• Suite-owned hooks follow the existing hook-context ownership, so output from a suite's beforeEach() is attributed to that suite.
• If the source location of an attributed test or suite is unavailable, file falls back to the child's entry file and remains a string.

Tests

• test/parallel/test-runner-console-output-attribution.mjs
• all test/parallel/test-runner-* tests
• make test-only
• make lint
• make test-doc

Fixes: #53103

[nodejs-github-bot]

Review requested:

• @nodejs/test_runner

[codecov]

Codecov Report

❌ Patch coverage is 96.11650% with 4 lines in your changes missing coverage. Please review.
✅ Project coverage is 90.32%. Comparing base (c59cd6b) to head (d5a88d0).
⚠️ Report is 8 commits behind head on main.

Files with missing lines	Patch %	Lines
lib/internal/console/constructor.js	86.20%	4 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #65147      +/-   ##
==========================================
+ Coverage   90.31%   90.32%   +0.01%     
==========================================
  Files         759      759              
  Lines      248301   248425     +124     
  Branches    46864    46887      +23     
==========================================
+ Hits       224246   224394     +148     
- Misses      15471    15475       +4     
+ Partials     8584     8556      -28     

Files with missing lines	Coverage Δ
lib/internal/test_runner/harness.js	89.22% <100.00%> (+1.08%)	⬆️
lib/internal/test_runner/runner.js	94.45% <100.00%> (+0.01%)	⬆️
lib/internal/test_runner/tests_stream.js	93.67% <100.00%> (+0.71%)	⬆️
lib/internal/console/constructor.js	97.71% <86.20%> (-0.48%)	⬇️

... and 36 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[MoLow]

IMHO #64389 is a better approach rather than monkeypatching existing APis