Large Library-Based Regression Test Cases Minimization for Object-Oriented Programs

摘要

Large library-based regression test cases minimization technique for object-oriented programs has been depicted in this paper. These works have been carried out in three steps. In the first step, the original program is instrumented and executed with test cases. Library is made on the basis of these test cases, coverage of codes, and then the program is modified. In the second step, the modified program is analyzed by latent semantic analysis. It is making the latent semantic matches automatically between users given values and linear combination of its small text objects or variables or database of the software. Therefore, modified code is recorded by latent semantic analysis. Data flow sensitivity and context sensitivity are used for statically and dynamically analyzing the affected and unaffected objects along with the recorded modified codes. After precision data flow analysis, test cases are generated from affected objects with same test cases coverage and affected objects with new test cases coverage. Therefore, redundant test cases are reduced by new optimal page replacement algorithm and updated the library along with code coverage records. In the third step, the test cases of former and modified program are collected and sent to the test cases repository. Now the new optimal page replacement algorithm is implemented on the test cases repository and reduced the regression test suites. An Illustrative example has been presented to establish the effectiveness of our methodology.