Improved Modified Condition/ Decision Coverage using Code Transformation Techniques

摘要

Modified Condition / Decision Coverage (MC / DC) is a white box testing criteria aiming to prove that all conditions involved in a predicate can influence the predicate value in the desired way. In regulated domains such as aerospace and safety critical domains, software quality assurance is subjected to strict regulations such as the DO-178B standard. Though MC/DC is a standard coverage criterion, existing automated test data generation approaches like CONCOLIC testing do not support MC/DC. To address this issue we present an automated approach to generate test data that helps to achieve an increase in MC/DC coverage of a program under test. We use code transformation techniques for transforming program. This transformed program is inserted into the CREST TOOL. It drives CREST TOOL to generate test suite and increase the MC/DC coverage. Our tech-nique helps to achieve a signicant increase in MC/DC coverage as compared to traditional CONCOLIC testings. Our experimental results show that the proposed approach helps to achieve on the average approximately 20.194 % for Program Code Transformer(PCT) and 25.447 % for Exclusive-Nor Code Transformer. The average time taken for seventeen programs is 6.89950 seconds.