fix: prevent tsconfig cache pollution with separate raw and built caches

[Boshen]

Summary

• Fixes tsconfig cache mixes root: true and root: false causing incorrect resolution #963
• Adds should_build flag to TsConfig to control path normalization separately from the root flag
• Splits the tsconfig cache into tsconfigs_raw (after callback, before build) and tsconfigs_built (after callback and build)
• For root=true (resolution): uses built cache
• For root=false (extending): uses raw cache

Test plan

• All existing tests pass (156 tests)
• The 3 new cache pollution tests from tsconfig cache mixes root: true and root: false causing incorrect resolution #963 pass
• just lint passes
• just fmt passes

🤖 Generated with Claude Code

[codecov]

Codecov Report

❌ Patch coverage is 88.23529% with 4 lines in your changes missing coverage. Please review.
✅ Project coverage is 94.17%. Comparing base (f6fbb23) to head (941cb0f).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/tsconfig.rs	57.14%	3 Missing ⚠️
src/cache/cache_impl.rs	96.29%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #970      +/-   ##
==========================================
- Coverage   94.36%   94.17%   -0.20%     
==========================================
  Files          17       17              
  Lines        3321     3346      +25     
==========================================
+ Hits         3134     3151      +17     
- Misses        187      195       +8     

☔ 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]

CodSpeed Performance Report

Merging this PR will improve performance by 4.38%

_{Comparing fix-tsconfig-cache-pollution (941cb0f) with main (f6fbb23)}

Summary

⚡ 1 improved benchmark
✅ 11 untouched benchmarks
⏩ 5 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	complex_real	22.9 µs	21.9 µs	+4.38%

Footnotes

1. 5 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[sapphi-red]

@Boshen I think you want to remove this as it may cause flaky failures.