QCEngine Windows Unicode Fix

[loriab]

Description

I started having trouble with a Psi4 test case (tests/nbody-multi-leve/) at psi4/psi4#3389 only on Windows and having to do with the § symbol. I'm not sure about the timing but perhaps something changed in GHA Win CI? I've compensated for it in psi4 with envvar PYTHONIOENCODING=utf-8, but AI (Haiku 4.5) suggests this is the more robust sol'n.

I've modified /psi/gits/QCEngine_next/qcengine/util.py to add robust error handling for subprocess output decoding on Windows.

The Fix (2-line change):

• Line 406: Changed stdout.getvalue().decode() → stdout.getvalue().decode(errors='replace')
• Line 407: Changed stderr.getvalue().decode() → stderr.getvalue().decode(errors='replace')

Why This is the Right Upstream Fix:

1. Root cause location: The popen() function is where qcengine actually decodes subprocess output - it's the
   central point of failure
2. Robust and safe: Using errors='replace' is the Python standard for handling uncertain encodings. It:
   - Converts invalid UTF-8 sequences to U+FFFD (replacement character)
   - Never crashes, just replaces bad bytes
   - Is backward compatible - valid UTF-8 works exactly the same
3. Comprehensive: This fix applies to ALL qcengine subprocess calls, not just Psi4
4. Minimal: Only adds an error handling parameter - no logic changes

Changelog description

Status

• Code base linted
• Ready to go

[codecov]

Codecov Report

❌ Patch coverage is 17.39130% with 19 lines in your changes missing coverage. Please review.
✅ Project coverage is 26.92%. Comparing base (cdda366) to head (8169b0b).
⚠️ Report is 26 commits behind head on master.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

This PR hardens QCEngine’s subprocess output capture on Windows by preventing UnicodeDecodeError when decoding stdout/stderr, addressing failures triggered by non-UTF-8 bytes (e.g., Windows code pages).

Changes:

• Decode captured stdout/stderr with errors="replace" to avoid crashing on invalid byte sequences.
• Add inline commentary explaining the Windows/non-UTF-8 motivation for the change.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.