How should you handle a crash during testing?

• A. Do not reproduce the crash to avoid bias.
• B. Reproduce the crash, capture logs and crash reports, document steps to reproduce, and report with frequency and conditions.
• C. Ignore intermittent crashes and continue testing.
• D. Only record screenshots without logs.

Answer: (B)

Reproducing a crash and gathering full evidence is essential for diagnosing and fixing defects effectively. When a crash occurs, you want to recreate it reliably so you can observe the exact sequence of actions, the conditions that trigger it, and how the system behaves under those circumstances. The best approach is to reproduce the crash, capture logs and crash reports, document steps to reproduce, and report with frequency and conditions. This provides concrete data: the build and environment, the data used, the precise steps, how often it happens (including intermittent behavior), and the accompanying error messages or stack traces. With this information, developers can pinpoint the root cause, verify that a fix works, and prevent the issue from reappearing in future builds. Other options fall short because they withhold critical details—without logs, reports, and a clear reproduction path you can’t reliably diagnose or verify a fix. Recording only screenshots, for example, omits the technical context needed to understand what went wrong.