gh-workflow-mcp: Additional test coverage for edge cases and integration

Additional Test Coverage for Edge Cases and Integration

Tracking 6 test coverage improvements identified in PR #432 all-hands-review (iteration 3) but deemed out-of-scope for immediate implementation.

Items from Test Analyzer

Issue #6: End-to-End Monitor Function Tests

Priority: MEDIUM
Files: monitor-run.ts, monitor-pr-checks.ts

Currently missing true end-to-end tests that:

• Start a real GitHub Actions workflow (in test environment)
• Monitor it with actual gh CLI commands
• Verify completion detection and status reporting
• Test both success and failure paths

Rationale: Current integration tests mock gh CLI. E2E tests would catch issues with actual GitHub API behavior, rate limits, and timing edge cases.

Implementation suggestion:

• Create test repository with simple workflow
• Use GitHub API to trigger runs
• Test monitor functions against real runs
• Add to CI but with appropriate timeouts and retry logic

Issue #7: Actual Timeout Behavior Tests

Priority: MEDIUM
Files: gh-watch-integration.test.ts

Current tests mock delays but don't verify:

• Actual AbortController signal propagation
• Timer cleanup on abort vs natural completion
• Race conditions between timeout and completion
• Signal handling in execa subprocess

Rationale: Mocked tests verify logic but not actual Node.js timeout mechanisms.

Implementation suggestion:

• Create tests that actually wait for timeouts (with short durations)
• Verify no hanging promises or unclosed timers
• Test timer cleanup in error paths
• Use --timeout in test runner for safety

Additional Edge Cases (4 items)

Watch command stderr output handling

Test createAbortableWatch with various stderr patterns:

• Warning messages mixed with success
• Progress indicators during long runs
• Unexpected output formats

Multiple concurrent run monitoring

Test monitor-run.ts with:

• Runs completing in different orders
• Some runs timeout, others succeed
• Simultaneous failure detection across runs

PR checks with changing status

Test monitor-pr-checks.ts with:

• Checks added after monitoring starts
• Required vs optional checks
• Check re-runs during monitoring

Fail-fast race conditions

Test edge cases in fail-fast mode:

• Failure detected exactly when watch completes
• Multiple simultaneous failures
• Polling errors during failure detection

Priority

Medium - Would improve confidence in edge cases but core functionality is well-tested (122+ tests passing).

Related Issues

• #432 - Parent PR for gh watch implementation
• #15372 - Documentation improvements (formerly planned as #540)

Implementation Notes

• E2E tests should be optional or have reasonable timeouts
• Consider test environment configuration (test repo, credentials)
• Balance coverage vs test execution time
• May need separate test suite for integration tests