Automated Regression Testing for Xianyu Transaction Chain

Background: Xianyu transaction chain is a critical link with multiple businesses, states, and operations. Order operations vary by type and status.

Problems: (1) Wide business scope makes impact assessment hard; (2) Need to support both new and old link versions; (3) Over 100 order tags require fast validation.

Solution: Use interface traffic recording/playback and scenario link verification to build an automated test task set. Tests run in the release pipeline; a checkpoint validation decides whether to proceed.

Automated test set design: Select cases that maximize automation benefit, ensure predictability of results, and cover core transaction scenarios.

Data preparation: Generate test data (product, user, funds) via the Test Data Construction Platform, separating data from test cases.

Case writing guidelines: • Reasonable decomposition of complex scenarios; • Simplify cases by reusing previous node data; • Multi‑layer assertions covering contracts, order data, business rules; • Reflect real user end‑to‑end flows.

Pipeline checkpoints: After test cases are settled, they are linked to the Aone release pipeline. The pipeline automatically triggers tests after deployment; if checkpoint validation fails, the release is blocked (strong checkpoint) requiring special approval.

Summary: Automated regression now validates core transaction interfaces and C2C scenarios, reducing manual verification from 30 minutes to 1 minute, improving quality assurance and release safety. Future work will expand the test library to more core business scenarios.