Systematic Extraction of Tests from Object-Oriented Programs.

Existing program-based automated test techniques from object-oriented programs generate only test data or test cases, which are not equipped with effective oracle to reveal the logical errors in the program. In addition, these techniques often focus on conventional code coverage criteria and intra-method testing, and are less concerned with inter-method, intra-class and inter-class testing. In this paper, we propose an automated testing approach to cover the inter-method and intra-class test levels. This approach generates tests that are equipped with effective oracles in terms of expected outcomes to reveal logical errors in the program under test. In order to demonstrate the applicability of the proposed approach, we applied it to a case study containing 14 different classes implemented in Java. Furthermore, we created artificial faulty versions of our case study, and the proposed approach was able to extract tests that reveal failures in 74% of faulty cases.