process: expose enhanced stack trace to uncaughtException handlers

[santusht06]

When an EventEmitter instance emits an unhandled 'error' event, Node attaches an internal stack enhancer (kEnhanceStackBeforeInspector) capturing the call site where emit('error', ...) was invoked.

However, createOnGlobalUncaughtException() in lib/internal/process/execution.js was dispatching the error to uncaughtExceptionMonitor and uncaughtException listeners before executing fatalExceptionStackEnhancers.beforeInspector(er).

As a result, custom uncaught exception handlers and APM monitoring libraries were unable to access the enhanced call site on err.stack.

Changes

• Invoked fatalExceptionStackEnhancers.beforeInspector(er) in createOnGlobalUncaughtException() prior to emitting uncaughtExceptionMonitor and uncaughtException.
• Added test/parallel/test-process-uncaught-exception-enhanced-stack.js to verify that both uncaughtExceptionMonitor and uncaughtException receive the enhanced err.stack containing the emit('error', ...) call site for EventEmitter and subclass instances.

Fixes: #55838

[nodejs-github-bot]

Review requested:

• @nodejs/loaders

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.19%. Comparing base (8af7545) to head (9e69cf3).
⚠️ Report is 7 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #65580      +/-   ##
==========================================
+ Coverage   90.18%   90.19%   +0.01%     
==========================================
  Files         771      771              
  Lines      264619   264640      +21     
  Branches    50231    50238       +7     
==========================================
+ Hits       238653   238701      +48     
+ Misses      16961    16947      -14     
+ Partials     9005     8992      -13     

Files with missing lines	Coverage Δ
lib/internal/process/execution.js	100.00% <100.00%> (ø)

... and 24 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.

[santusht06]

Thank you @MikeMcC399. I have closed #65531 and #65543 to focus solely on this PR. I will fix the CI failures and run the full test suite before pushing any further updates. Apologies for not following the one-PR-at-a-time guideline.

[santusht06]

Rebased onto the latest upstream/main. The CI build failures (shared libraries on aarch64/x86_64) are pre-existing infrastructure issues unrelated to this change — this PR touches only two JavaScript files (lib/internal/process/execution.js and the new test). All process and EventEmitter tests pass locally.

[santusht06]

Converting to draft. After deeper analysis I found a critical bug in the current approach:

fatalExceptionStackEnhancers.beforeInspector() mutates er.stack in-place but does not remove the kEnhanceStackBeforeInspector symbol from the error object. In the no-handler path (when the process exits), the C++ layer (ReportFatalException in src/node_errors.cc) independently calls enhance_fatal_stack_before_inspector — which resolves to the same beforeInspector function. This causes enhanceStackTrace() to run twice, doubling the 'Emitted error event at:'\ frame in the crash output for every unhandled EventEmitter error. This would be a visible regression for all Node.js users not using uncaughtException` handlers.

Need to design this more carefully before proceeding.