Enhancement: Improve test suite performance

[dougborg]

Summary

The test suite currently takes ~13 minutes to run locally (803 seconds for 250 tests), which impacts developer productivity and CI feedback loops.

Current Performance

• Local execution: 13 minutes 23 seconds (803.49s)
• 250 tests (11 xfailed)
• CI execution: ~2-3 minutes per Python version × 5 versions = 10-15 minutes total

Potential Optimizations to Investigate

1. Parallel Test Execution

• Use to run tests in parallel
• Could reduce runtime by 2-4x depending on CPU cores
• Command: pytest -n auto

2. Test Fixtures Optimization

• Review fixture scope (function vs module vs session)
• Cache expensive setup operations (file generation, parsing)
• Lazy load test data

3. Selective Test Running

• Implement test markers for fast/slow tests
• Run quick smoke tests first in CI
• Full suite only on main branch or nightly

4. Code Generation Caching

• Cache generated code for identical OpenAPI specs
• Reduce redundant generation in similar test cases

5. CI-Specific Optimizations

• Matrix strategy to fail-fast on first failure
• Split test suite across multiple jobs
• Use GitHub Actions cache for dependencies

Success Criteria

• Reduce local test runtime to < 5 minutes
• Reduce CI feedback time to < 2 minutes per Python version
• Maintain 90%+ code coverage
• No flaky tests introduced

Non-Goals

• Don't sacrifice test coverage for speed
• Don't introduce test interdependencies
• Don't skip important integration tests

Priority

Medium - Developer quality of life improvement, not blocking

---

Discovered during #98 refactor work where full test suite verification took significant time.