BugReplay

Bug reports often describe a symptom without providing enough evidence to reproduce or verify the problem. This creates unnecessary back-and-forth between developers, testers, and reviewers. A developer may fix the visible issue, but without a reliable reproduction and regression test, the same bug can easily return later. BugReplay is a small evidence-first debugging prototype built around this problem. It demonstrates a fictional checkout defect where the SAVE10 coupon is applied twice to a $50 cart. The correct result after one 10% discount is $45, but the original buggy implementation returns $40 because the same discount is subtracted twice. The prototype allows a user to enter notebook and pen-set quantities, provide a coupon code, choose between the original bug and repaired code, and calculate the checkout total. The original variant reproduces the incorrect result, while the repaired variant applies the discount once and returns the expected total. The verification page then presents the before-and-after evidence together with the regression-test outcome. The workflow is intentionally simple: report the problem, reproduce it with controlled inputs, inspect the calculation, apply a focused repair, add a regression assertion, and verify the result. This makes the debugging process easier to understand and review. BugReplay is intended for developers and reviewers who need a clear connection between a bug report, the code change, and the proof that the change works. It uses only fictional synthetic data. There is no real payment processing, customer information, authentication, or production integration. The project is built with Next.js, React, TypeScript, Vitest, and IBM Bob IDE. It is deployed as a working prototype on Vercel. GitHub repository: https://github.com/Kunal435573/bugreplay Live demo: https://bugreplay.vercel.app/bugs/BUG-001 Verification page: https://bugreplay.vercel.app/bugs/BUG-001/verification