QA Transformation: Applying HTTP DIFF and Visual UI Automation to Operational and Order‑Related Requirements

With the industry’s rapid development, QA is no longer limited to functional testing for code quality or release assurance; it now explores technical means to assist developers in self‑testing, moving from a traditional role to a more efficient, supportive one. Under this backdrop, the ZuanZuan YouPin QA team began exploring transformation methods.

Transformation Path

During the exploration, the team realized that testing tools and technical solutions must be targeted rather than applied blindly. Different requirement types demand specific techniques to truly improve efficiency and provide assistance. The article introduces two distinct requirement categories and the corresponding technologies employed.

1. Operational Requirements

YouPin’s B2C business relies heavily on user‑facing operational activities, which typically feature short business flows, frequent page changes, and few modification points. These characteristics make them an ideal entry point for transformation.

Two tools were explored for this category:

HTTP DIFF

HTTP DIFF is suitable for short‑flow, interface‑heavy scenarios and for regression testing of existing functionality. It was adopted as the first tool for operational needs.

Sample screenshots of the tool’s interface (interface list, new interface, diff result) are shown below.

Visual UI Automation

While HTTP DIFF validates the API layer, many operational activities also require front‑end visual verification. Traditional UI automation tools (Appium, Puppeteer) are unsuitable for highly dynamic pages due to high maintenance costs. Instead, a visual‑based approach captures screenshots of target pages/areas, aggregates them, and lets humans perform semi‑automatic verification, offering resilience to layout changes and finer detail checks.

These tools together address the operational requirement transformation.

2. Order‑Related Requirements

Order processes involve monetary transactions and thus remain heavily QA‑driven. Two key challenges are rapid order/data construction during testing and comprehensive coverage of complex billing logic.

Data Construction

Data construction combines multiple API calls into a single RPC or HTTP endpoint, enabling testers to obtain ready‑made data for any required node, thereby simplifying test data setup.

Online Order Inspection

Full regression of complex order logic is costly. The team proposes a scheduled online inspection that periodically captures diverse live orders and validates each monetary logic point, serving as a safety net for regression. This tool is still in design and exploration.

Conclusion

The transformation journey has just begun; several tools have been built and ideas generated, but the critical task now is to promote tool adoption and practical deployment. Keeping pace with the inevitable transformation wave is essential for QA professionals to thrive in future testing models.