[rozzilla]

Use indexed writes instead of primordial push calls when collecting dependent signals to abort.
Abort-throughput benchmarks on Apple M2/macOS 15.7.4, using rebuilt Node v28.0.0-pre binaries against 47438d9. Fixtures were created outside timing, with 30 samples per binary:

Dependents per source	Aborts per sample	Candidate vs baseline
16	2,000	+18.55% ±11.94%
64	1,000	+28.07% ±10.37%

Validation:

• make -j8: passed
• make lint-js: passed
• python3 tools/test.py parallel/test-abortsignal-any.mjs parallel/test-abortsignal-drop-settled-signals.mjs: passed
• make -j4 test: not run

[nodejs-github-bot]

Review requested:

• @nodejs/performance

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.68%. Comparing base (47438d9) to head (afd1bf6).
⚠️ Report is 3 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #67299      +/-   ##
==========================================
- Coverage   93.00%   90.68%   -2.33%     
==========================================
  Files         428      797     +369     
  Lines      195814   279609   +83795     
  Branches    30287    53780   +23493     
==========================================
+ Hits       182117   253557   +71440     
- Misses      13371    16347    +2976     
- Partials      326     9705    +9379     

Files with missing lines	Coverage Δ
lib/internal/abort_controller.js	95.41% <100.00%> (-0.01%)	⬇️

... and 499 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.