testing-strategy.adoc

Chronicle-Algorithms Testing Strategy

Table of Contents

• 1. Scope
• 2. Unit and Property Testing
• 3. Fuzz and Negative Testing
• 4. Performance Regression
• 5. Static Analysis and Checks
• 6. Observability Expectations

1. Scope

Testing guidance for hashing, bit set, byte access, and lock primitives in Chronicle-Algorithms.

2. Unit and Property Testing

• Hashing: verify avalanche/distribution properties with randomised inputs and seeded property tests.

• Bit sets: cover empty, dense, and sparse cases; test boundary indices and concurrency assumptions where applicable.

• Bytes accessors: validate bounds checks, slice/offset handling, and behaviour with on-heap vs direct vs Chronicle Bytes backings.

• Locks: assert state transitions and contention paths without relying on java.util.concurrent locks.

3. Fuzz and Negative Testing

• Feed random and malformed inputs through hashing and bounds-sensitive paths to catch overflows and index errors.

• Include adversarial cases (large lengths, near-overflow sizes) for off-heap friendly accessors.

4. Performance Regression

• Maintain microbenchmarks (e.g., JMH) for hot paths: hash computation on typical payload sizes, bit set set/get/scan operations.

• Track baseline metrics; flag regressions when latency/throughput deviates beyond agreed thresholds.

5. Static Analysis and Checks

• Keep Checkstyle and SpotBugs clean; document any suppression rationale inline.

• Prefer deterministic tests with fixed seeds; log seeds on failure when randomness is involved.

6. Observability Expectations

• No hidden threads or background tasks in tests; keep measurements isolated.

• Fail fast on misuse (bounds violations) rather than silently truncating or wrapping inputs.