fix: resolve NamedTemporaryFile permission errors on Windows in tests

[Lampard7crypt]

No description provided.

[cffls]

#479 should be able to resolve issue #478 . Could you please explain if there is any other issue this PR tries to address?

[Lampard7crypt]

#479 should be able to resolve issue #478 . Could you please explain if there is any other issue this PR tries to address?

1. decode_arrayinfinite recursion (serialization.py)
   PyCardano overrides cbor2's array decoder to preserve indefinite-length CBOR arrays as IndefiniteFrozenList. Both branches of the override called self.decode_array() —which after patching major_decoders[4] re-entered the same function, causing infinite recursion. As a side effect, for arrays with 24+ items this double-called _decode_length on the stream, corrupting it and producing CBORDecodeEOF.

Fix: Save major_decoders[4] before patching, then delegate to that saved reference instead of calling self.decode_array(). This breaks the recursion loop entirely.

2. Windows NamedTemporaryFile permission errors (7 tests)
   What: On Windows, NamedTemporaryFile() holds an exclusive file lock while open. When save() tried to open the same path for writing, it got PermissionError. Works fine on Linux/macOS where locking is advisory.

Fix: Use delete=False so the lock is released immediately, then manually delete the file with os.unlink() in a finally block.

[cffls]

Fix: Save major_decoders[4] before patching, then delegate to that saved reference instead of calling self.decode_array(). This breaks the recursion loop entirely.

self.decode_array(subtype) resolves to CBORDecoder.decode_array, which is the original unpatched instance method on the class — it was never modified. Saving _original_decode_array = cbor2._decoder.major_decoders[4] before patching and calling it explicitly does the exact same thing, just with more indirection.

I am fine with the changes for the temp file usage though.

[Lampard7crypt]

Fix: Save major_decoders[4] before patching, then delegate to that saved reference instead of calling self.decode_array(). This breaks the recursion loop entirely.

self.decode_array(subtype) resolves to CBORDecoder.decode_array, which is the original unpatched instance method on the class — it was never modified. Saving _original_decode_array = cbor2._decoder.major_decoders[4] before patching and calling it explicitly does the exact same thing, just with more indirection.

I am fine with the changes for the temp file usage though.

Oh, thanks for the clarification, you're right that self.decode_array(subtype) resolves to the unpatched class method since we only patched the major_decoders dict and not the class itself, so saving the explicit reference is indeed redundant. I've reverted those changes. The NamedTemporaryFile fixes I've kept as-is.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.60%. Comparing base (1d2e825) to head (907cd8b).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #485      +/-   ##
==========================================
+ Coverage   90.57%   90.60%   +0.03%     
==========================================
  Files          34       34              
  Lines        5154     5154              
  Branches      781      781              
==========================================
+ Hits         4668     4670       +2     
+ Misses        306      305       -1     
+ Partials      180      179       -1     

☔ 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.

[cffls]

lgtm, thanks!