zlib: reject ambiguous ZIP archive ends

[mcollina]

ZIP archives can contain more than one structurally plausible EOCD record. The previous exact-comment-first search could choose a different central directory depending on trailing padding.

Scan candidates in the same tail window used for file-backed archives and reject multiple plausible interpretations. Invalid EOCD-like comment bytes remain ignored, and padded archives with one interpretation remain supported.

[pipobscure]

I like the fix since it both matches the ecosystem, but still keeps our slightly tighter strictness. There are a few cases where exotic archives would be rejected for being wrong failing closed.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.28%. Comparing base (5ba72ae) to head (48a3069).
⚠️ Report is 70 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #65007 +/- ##
==========================================
- Coverage 90.29% 90.28% -0.01% 
==========================================
Files 760 762 +2 Lines 247061 247699 +638 Branches 46586 46718 +132 ==========================================
+ Hits 223076 223642 +566 - Misses 15448 15489 +41 - Partials 8537 8568 +31 

Files with missing lines	Coverage Δ
lib/internal/zip/headers.js	97.51% <100.00%> (+0.25%)	⬆️

... and 70 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.