Building a C++ Cross‑Platform Framework for WeChat Pay: Architecture, Routing, and Quality Assurance

WeChat Pay faces significant challenges across different operating systems and applications, primarily due to fragmented platform implementations that cause bugs, poor scalability, inconsistent quality assurance, and divergent user experiences.

To address these core issues, a C++‑based cross‑platform framework was built, consolidating the payment flow for iOS (from version 7.0.4) and Android (from version 7.0.7). The new architecture reduced the core payment code by nearly 45% (3512 lines vs. 6328 lines native) and accelerated feature delivery, completing version 7.0.4 requirements in just three person‑days.

Key architectural steps:

Abstract business processes into UseCase components, separating flow logic from UI controllers.

Introduce a routing mechanism that encapsulates payment‑related actions (e.g., opening a WebView, launching a mini‑program, showing a dialog) and passes a unified data model.

Refactor network requests: each CGI becomes an independent command object with a clear lifecycle tied to its initiating UseCase , eliminating one‑to‑many communication bugs and lifecycle mismatches.

Standardize data transmission by removing shared mutable models, using immutable value types, and enforcing unidirectional data flow with dependency injection and delegate patterns for necessary callbacks.

These changes yielded measurable improvements: crash rates remained stable after launch, performance gains were evident from the reduced codebase, and special‑process handling (WebView, mini‑program, dialogs) saw an 83% code reduction.

The evolved architecture now consists of three layers: (1) abstracted business flow, (2) routing‑driven navigation, and (3) disciplined network request management, all built on a clean C++ core that can be shared across iOS and Android.

In summary, the article argues that software architecture’s purpose is to manage complexity; by iteratively refining the WeChat Pay client architecture, the team achieved higher reliability, maintainability, and productivity, laying a foundation for future feature development and automated data reporting.