iamHrithikRaj C++ reliability
Ownership and lifetime
- Identify the owner of every relevant object and resource.
- Check whether references, pointers, spans, views, callbacks, and captures can outlive their targets.
- Treat shared ownership as a deliberate design decision, not a default.
- Verify destruction ordering and shutdown behavior.
- Check move operations for valid source and destination states.
Resource safety
- Prefer RAII for memory, handles, locks, registrations, and transactions.
- Verify cleanup after partial initialization and every error path.
- Check exception guarantees where exceptions are enabled.
- Ensure destructors cannot unexpectedly propagate failures.
Concurrency
- Document shared state and its synchronization.
- Check lock ordering, callback reentrancy, races, deadlocks, and cancellation.
- Verify that atomics use an ordering strong enough for the invariant.
- Examine object lifetime across asynchronous work.
Error handling
- Preserve actionable error context.
- Avoid broad catches and success-shaped fallbacks.
- Verify that failed operations leave state valid and recoverable.
- Check retry safety, idempotency, and duplicate effects.
Performance
- Look for avoidable allocation, copying, contention, and cache-unfriendly access only when they are material to the workload.
- Require measurement before recommending complexity.
- Include correctness and operational costs in every optimization tradeoff.
Findings
Report only issues with a concrete failure scenario or evidence-backed material risk. Include severity, location, scenario, evidence, correction, and confidence.