performance-review

To check code changes for performance regressions, evaluate diffs across startup, frames, memory, energy, network, storage, and bundle size. This requires quantified measurements with device or build context, comparing results against documented budgets and baselines to identify risks.

A performance review is a measurement-driven evaluation of code diffs against predefined budgets. You need it during code reviews, UI implementation, test harnesses, and release-readiness checks to prevent costly regressions in startup time, frame rate, memory usage, and battery consumption.

To enforce performance budgets for memory usage and bundle size, apply bounded resource usage rules and cache invalidation requirements to the code diff. Budget violations are treated as blockers, resulting in a FAIL verdict until the metrics align with the established baseline.

Yes, a performance review requires quantified measurements that include explicit device or environment and build context. It also requires documented performance budgets, baseline comparisons, and optional platform auditor results to issue a valid PASS, PASS WITH NOTES, or FAIL verdict.

The best way to gate a release using performance budget violations is to treat them as blockers during the review process. Evaluate the diff against documented baselines and issue an explicit verdict of PASS, PASS WITH NOTES, or FAIL to ensure release-readiness checks are met.