Boost Test Efficiency: Smart Code Change Impact Evaluation and Automated Case Recommendation

Background Introduction

When a large batch of code changes is released, it is unclear which functionalities are affected and which rely on experience.

After testing, it is uncertain whether all impacted functional points have been covered.

In fast agile iterations with limited testing time, we must identify core traffic and ensure high availability for critical functions.

Obsolete code remains in services without clear purpose, and removing it is risky.

In a micro‑service architecture, evaluating the impact on business parties is challenging.

Solution Design

The core problems are test efficiency and test quality. We address them through the following steps:

Scientifically evaluate which functional points are affected by code changes and ideally provide intelligent test case recommendations.

Obtain incremental coverage to clearly see which changed code has not yet been covered.

Achieving these two points resolves the questions of what to test before testing and which test points to supplement during testing.

Test Functionality Evaluation

Key capabilities involved:

1. Collect method set from code changes

Use JGit to obtain code diffs.

Use classParser to parse class files and retrieve the constant pool.

2. Derive impacted interfaces from changed methods

BCEL: Analyze and manipulate class files by delving into JVM bytecode.

Constant pool: Contains all information in a class file; analyzing it reveals call relationships between functions.

Coverage Result Analysis

Based on incremental coverage, analyzing uncovered methods reveals which interfaces are affected, enabling targeted testing.

Interface Analysis

With the above capabilities we can identify which interfaces to test for a code change. Since interface importance varies, we need the following information to prioritize under limited time:

Interface priority to determine if it lies on a core functional path.

Whether the interface has automation coverage (including automated tests and traffic replay) and the method tree covered by automated cases.

Front‑end entry point for quick test access.

Business parties impacted, to decide if they need regression notifications.

Interface Priority Analysis

Select a one‑month window to fetch system API call statistics. Too long a window may include deprecated high‑traffic APIs; too short may be skewed by temporary events.

Fit the discrete data with a simple least‑squares method.

Identify inflection points on the fitted curve, assuming data with the same priority share a distribution.

Split the data at the inflection points to obtain priority intervals.

Interface Automation Coverage

Automation coverage can be retrieved from continuous integration results, requiring the automation platform to expose relevant APIs.

Automation Case Coverage of Code Functions

By integrating the automation framework with a precise testing and coverage platform, we can obtain the method tree covered by each automated case.

Interface Entry and Impacted Business Parties

Analyze service‑to‑service call data from the monitoring system to identify dependent interfaces and their entry points.

Intelligent Test Case Recommendation

Application Scenarios

With these capabilities, precise testing addresses the following problems:

Clearly define the impact scope of code changes, avoiding full‑scale testing.

The test case platform can intelligently recommend cases based on code changes, facilitating developer self‑testing; combined with incremental coverage, it raises the test entry threshold.

During single‑case testing, the collected method tree helps newcomers quickly understand the business.

Incremental coverage in integration testing evaluates test effectiveness and serves as a release gate.

Single‑case coverage data enables traffic reduction, improving regression efficiency.

Online coverage data marks redundant code.

Mid‑platform services can identify affected dependents, providing business teams with clear regression scopes.