Harden Windows queue Planfile probes against non-UTF8 subprocess output

[Copilot]

Queue mode still crashed on Windows with UnicodeDecodeError because some Planfile probe subprocesses were run in text mode, so Python reader threads decoded output as strict UTF-8 before Koru’s fallback logic could run. Ticket execution could complete, but probe paths still emitted background decode failures.

• Probe subprocess decoding moved to byte-safe path

  • In src/koru/queue/ticket.py, switched probe calls from subprocess.run(..., text=True) to text=False in:
    • _python_has_planfile_cli
    • _planfile_supports_structured_queue_json
    • _configured_planfile_cmd_usable
  • Decoding is now routed through koru.queue.runners._decode_subprocess_output, aligning probe behavior with queue runner robustness.

• Version / marker parsing now operates on safely decoded text

  • Structured-version detection and module-missing marker checks now parse decoded strings produced by the shared tolerant decoder, preventing code-page-specific byte sequences from crashing probe logic.

• Regression coverage for non-UTF8 probe output

  • Added focused tests in tests/test_planfile_queue.py for:
    • non-UTF8 bytes in version probe output (including old/new version behavior)
    • non-UTF8 bytes in configured-command probe output on both failure and success paths

# Before (can crash in reader thread on Windows code page output)
subprocess.run(cmd, capture_output=True, text=True, ...)

# After (capture bytes, decode with queue-safe fallback)
proc = subprocess.run(cmd, capture_output=True, text=False, ...)
text = _decode_subprocess_output(proc.stdout)

[tom-sapletta-com]

@copilot resolve the merge conflicts in this pull request

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.