During development and support phases, software is modified to enhance its functionality, detect faults, and adapt it to different platforms. Regression testing is used to identify faults that were introduced when modifying code [1-18] or to assure that a change, for example a functional enhancement, bug fix, patches or configuration changes, did not introduce new faults. However, a lack of test inputs that exercise a changer behavior is a common issue in regression testing. A large number of test inputs is generated in order to cover modified parts of the code. Then the tests are executed using generated test inputs on the old and new versions of the code, differences are identified and presented to developer with the details regarding the lines of changed code and the differences [2]. The proposed approach can provide developers with detailed information regarding code coverage and various statistics.
展开▼