worker: fix TOCTOU race in CWD caching

[mcollina]

The atomic counter used to signal CWD changes to worker threads was
being incremented before chdir() completed, creating a race window
where workers could cache stale directory paths with the new counter
value. This caused process.cwd() in workers to return incorrect values
until the next chdir() call.

Fix by reordering operations: call originalChdir() first, then
increment the counter. This ensures workers never cache stale data
while believing it is current.

A unit test for this fix is not feasible as it would be too flaky due to the timing-dependent nature of the race condition.

Reported-by: Giulio Comi
Reported-by: Caleb Everett

[mcollina]

cc @nodejs/tsc

[Renegade334]

FYI, a PR was opened by the community reporter at #61664.

[anonrig]

Nice catch.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.75%. Comparing base (81e05e1) to head (4b045af).
⚠️ Report is 357 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #61725      +/-   ##
==========================================
+ Coverage   88.53%   89.75%   +1.21%     
==========================================
  Files         703      675      -28     
  Lines      208538   204525    -4013     
  Branches    40224    39306     -918     
==========================================
- Hits       184629   183565    -1064     
+ Misses      15912    13241    -2671     
+ Partials     7997     7719     -278     

Files with missing lines	Coverage Δ
lib/internal/worker.js	96.52% <100.00%> (ø)

... and 199 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.