(WIP) Synthetic tests for zip64 extra_fields

[frheault]

Extra verification for zip64 extra_fields due to benchmarking across languages. Data from trx-rs seems to write each individual files as "large_file" which add the information in extra_fields. New test for safety.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 62.15%. Comparing base (951e7c9) to head (6f36d9c).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #115      +/-   ##
==========================================
+ Coverage   61.64%   62.15%   +0.50%     
==========================================
  Files          13       13              
  Lines        2623     2658      +35     
==========================================
+ Hits         1617     1652      +35     
  Misses       1006     1006              

Flag	Coverage Δ
unittests	62.15% <100.00%> (+0.50%)	⬆️

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

☔ View full report in Codecov by Harness.
📢 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.

[arokem]

Is this still WIP? It looks good to me.

[frheault]

@arokem Luckily (or not) the trx-python was the only implementation that required no fix for my interoperability tests and benchmark (you can go see https://github.com/tee-ar-ex/trx-manuscript-2026-benchmark/pull/1/changes#diff-5340cc3c2b57b554bf3988166c763967264c39bcc2f17e6a7bf9efa14ccae9a6) .

But I still wanted to have a PR per language in case it requires changes later. Also turns out the Python implentation of TRX being 90% numpy and so being C under the hood is the fastest?! (Still unsure about that, but I read it to memory and cast position to float32 to be as fair to the other as possible)