fix: TsconfigPathsPlugin circular references, unscoped pkg extends, and perf

[xiaoxiaojx]

Summary

Fix multiple bugs and performance issues in TsconfigPathsPlugin:

• Circular project references → stack overflow: When project references form a cycle (A references B, B references A) with references: "auto", _loadTsconfigReferences recurses infinitely. Fixed by tracking visited reference config paths in a Set and skipping already-visited ones.
• Unscoped npm package extends not resolved: extends: "my-base-config" (unscoped package, no sub-path) failed to resolve to node_modules/my-base-config/tsconfig.json. The existing fallback logic only handled scoped packages (@scope/name) and sub-path packages (pkg/tsconfig). Added an else branch for unscoped packages.
• readFile used for existence check: _loadTsconfigFromExtends called readFile just to check if a path exists, discarding the buffer. Changed to stat which avoids reading the entire file content unnecessarily.
• substituteConfigDir regex on every call: The regex replacement ran even when paths contained no ${configDir}. Added a fast includes check to skip the regex when not needed.

What kind of change does this PR introduce?

fix, perf

Did you add tests for your changes?

Yes. Added 3 new test cases:

• test/tsconfig-paths.test.js: "should handle circular references without hanging or crashing"
• test/tsconfig-paths.test.js: "should resolve paths from a circular-referenced project"
• test/tsconfig-paths.test.js: "should resolve paths inherited from an unscoped npm package tsconfig"

Test fixtures added:

• test/fixtures/tsconfig-paths/references-circular/ — A↔B circular reference project
• test/fixtures/tsconfig-paths/extends-unscoped-pkg/ — unscoped npm package extends

Does this PR introduce a breaking change?

No.

If relevant, what needs to be documented once your changes are merged or what have you already documented?

No documentation changes needed.

Use of AI

This PR was authored with the assistance of Claude Code (Claude Opus 4.6). The bugs were identified through code review, test cases were written first to reproduce, then fixes were applied and verified.

[changeset-bot]

🦋 Changeset detected

Latest commit: 510978b

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
enhanced-resolve	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.56%. Comparing base (292aee5) to head (510978b).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #575      +/-   ##
==========================================
+ Coverage   96.54%   96.56%   +0.01%     
==========================================
  Files          50       50              
  Lines        2927     2937      +10     
  Branches      922      928       +6     
==========================================
+ Hits         2826     2836      +10     
  Misses         85       85              
  Partials       16       16              

Flag	Coverage Δ
integration	96.56% <100.00%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 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.

[codspeed-hq]

Merging this PR will degrade performance by 94.56%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 3 improved benchmarks
❌ 1 regressed benchmark
✅ 134 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Memory	resolve-to-context: directory resolve (warm)	96.6 KB	1.8 KB	×54
⚡	Memory	tsconfig-paths: 5 path prefixes (warm)	101.9 KB	1.8 KB	×58
❌	Memory	deep-hierarchy: relative from 10-deep dir (warm)	1 KB	18.6 KB	-94.56%
⚡	Simulation	cache-predicate: mixed cached/uncached requests (warm)	4 ms	2 ms	×2

---

_{Comparing fix/tsconfig-paths-plugin-bugs (510978b) with main (292aee5)}