Flaky test: test_spawn_with_safe_arguments fails sporadically on Windows

[logbie]

Bug Description

The test test_spawn_with_safe_arguments in tests\subprocess_security_test.rs is failing sporadically with the error "Output should contain expected text".

Error Details

thread 'test_spawn_with_safe_arguments' (10228) panicked at tests\subprocess_security_test.rs:204:5:
Output should contain expected text
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

failures:
    test_spawn_with_safe_arguments

test result: FAILED. 9 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.34s

error: test failed, to rerun pass `--test subprocess_security_test`

Environment

• WFL Version: 26.1.17
• Operating System: Windows 11
• Failure Pattern: Sporadic (intermittent failures)

Impact

This flaky test can cause CI/CD pipelines to fail intermittently and creates uncertainty about the actual health of the codebase. The test appears to be related to subprocess security functionality, so ensuring its reliability is important for validating WFL's security features.

Suggested Investigation

1. Add more detailed logging to understand what output is being produced vs. expected
2. Check for timing-related issues in subprocess spawning on Windows
3. Consider adding retry logic or more robust output validation
4. Run with RUST_BACKTRACE=1 to get more debugging information

[logbie]

@claude

Investigate and fix this it looks like a flaky test but i am not 100% sure

[claude]

Claude finished @logbie's task —— View job • claude/issue-239-20260109-1611 • Create PR ➔

---

Flaky Test Fixed Successfully ✅

I've investigated and fixed the flaky test_spawn_with_safe_arguments test that was failing sporadically on Windows.

Todo List:

• Read and analyze the failing test code
• Understand the assertion that's failing at line 204
• Add detailed logging to capture actual vs expected output
• Identify Windows-specific timing or subprocess issues
• Implement fixes for more robust output validation
• Test the changes
• Commit and push the fix

Root Cause Analysis:

The test was failing due to a race condition in the WFL subprocess handling:

1. Process Handle Cleanup: The wait_for_process function removes the process handle from the process map when called (line 1005-1006 in src/interpreter/mod.rs)
2. Timing Issue: The original test waited 200ms then tried to read output after waiting for process completion, but the process ID was already cleaned up
3. Windows Sensitivity: The timing was particularly sensitive on Windows systems where subprocess spawning can be slower

Fix Implementation:

Key Changes:

• Increased wait time from 200ms to 300ms to give processes more time to complete and output to be captured
• Reordered operations to read output BEFORE waiting for process completion, preventing the process handle from being cleaned up
• Better error messages that show actual vs expected output for easier debugging
• Applied same fix to test_multiple_safe_processes for consistency

File Changed: tests/subprocess_security_test.rs

The fix ensures that:

1. Subprocess output is captured before the process handle is removed
2. More generous timing accommodates slower Windows systems
3. Better error reporting helps diagnose future issues
4. All subprocess security tests continue to pass

Verification:

• ✅ All 10 subprocess security tests pass consistently
• ✅ Multiple test runs show stable behavior
• ✅ Fixed test now provides clear error messages if it fails

---