A unified framework for efficient test case generation and prioritization driven by multiple coverage criteria

Software testing is fundamental to system reliability and remains resource-intensive. Traditional frameworks often treat test case generation and prioritization as disconnected phases, neglecting the rich coverage metadata produced during generation and necessitating costly pre-execution. To bridge this gap, we present an integrated framework that enables profiling-free prioritization by reusing generation-time coverage metadata, with the objective of enabling earlier accumulation of multi-criteria structural coverage during test execution, and empirically facilitating earlier fault exposure. The proposed framework features Multiple Coverage Test Case Generation (MCTCG), an enhanced symbolic execution tool engineered to generate test suites satisfying six diverse coverage criteria while simultaneously capturing granular coverage data. Leveraging this intrinsic data, TCPall-first performs greedy multi criteria coverage prioritization by maximizing marginal structural coverage in the execution prefix, without requiring re-execution. Experimental results on CodeNet and RERS benchmarks demonstrate that MCTCG achieves superior coverage and bug detection rates compared to baseline tools. Furthermore, TCPall-first significantly outperforms state-of-the-art prioritization methods, improving the Average Percentage of Faults Detected (APFD) by 4.11% and reducing the first-fail metric by 8.7%. This work effectively enhances testing efficiency by seamlessly unifying the generation and prioritization processes, making it particularly valuable for standard-compliant white-box testing.