CampusLearn
← BIT 305 Mobile Application Development

Testing, Packaging and Publishing Apps

Notes

Testing real mobile conditions Mobile testing must cover application logic, interfaces and operating conditions. Unit tests check isolated behaviour, while instrumented tests run with Android framework support on a device or emulator. Test different screen sizes, supported system versions, slow networks, denied permissions and process recreation. Accessibility checks should include screen reader labels and navigation order. A release candidate should also be tested against realistic server behaviour rather than only mock responses. Crashes, duplicate submissions and data loss deserve explicit investigation because a visually correct interface can still fail during interruptions or unstable connectivity. Preparing and distributing releases Packaging turns application code and resources into a distributable artefact. An Android App Bundle supports store-generated packages for device configurations, while signing establishes release authenticity and continuity. Protect signing credentials and follow the chosen distribution channel requirements. Increase version identifiers for new releases, remove debug settings and verify production service endpoints. Publishing also requires accurate descriptions, privacy disclosures and permission declarations. A staged rollout releases to a limited audience first, allowing monitoring before wider distribution. Track crashes, failed requests and user feedback, and maintain a recovery plan rather than treating publication as the end of development. Practical example A Nairobi e-commerce startup has 5,000 eligible app users and begins a release with a 10 percent rollout, covering 500 users. Monitoring records eight crashes across 2,000 sessions, giving a crash-event rate of 8 divided by 2,000 times 100, or 0.4 percent. This is not automatically the percentage of users affected because one user may have several sessions or crashes. The team also finds six duplicate checkout requests and pauses expansion while investigating retry behaviour. After fixing the issue, it tests upgrade paths and database migrations before increasing distribution. A tested older build alone may not reverse an incompatible database change safely.

Quiz

1. Why is eight crashes in 2,000 sessions not necessarily the affected-user percentage?

2. What is the best response to duplicate checkout requests during a staged rollout?

3. A 10 percent rollout to 5,000 eligible users initially covers how many users?

Sign in to submit and securely grade your quiz.