%0 Journal Article
%T Regression Test Reduction for Object-Oriented Software: A Control Call Graph Based Technique and Associated Tool
%A Nicolas Frechette
%A Linda Badri
%A Mourad Badri
%J ISRN Software Engineering
%D 2013
%R 10.1155/2013/420394
%X This paper presents a selective regression testing technique and an associated tool for object-oriented software. The technique is based on the concept of Control Call Graphs, which are a reduced form of traditional Control Flow Graphs. It uses static analysis of the source code of the program. The developed tool (1) identifies the Control Call Paths potentially impacted by changes, (2) selects, from an existing test suite, the appropriate test cases, and (3) generates new JUnit test cases for control call paths that are not covered by existing tests (new ones, or those whose structure has been modified after changes). In this way, the approach supports an incremental update of the test suite. The selected JUnit test cases, including the new ones, are automatically executed. Three concrete case studies are reported to provide evidence of the feasibility of the approach and its benefits in terms of reduction of regression testing effort. 1. Introduction Software systems need to continually evolve for various reasons: adding new features to satisfy user requirements, changing business needs, introducing novel technologies, correcting faults, improving quality, and so forth. It is, therefore, important to ensure that modifications do not adversely affect the software. Software maintenance is an essential activity [1]. The costs of software maintenance keep, however, increasing [2]. Among the various maintenance activities, regression testing represents a crucial one. Regression testing is actually an important activity to ensure software quality, particularly when software is actively maintained and updated. It can also be used in the testing release phase of software development. Regression testing is the process that consists of determining if a modified software system still verifies its specification and whether new errors were introduced inadvertently [3¨C5]. For obvious reasons, the retest-all approach that consists in rerunning every test case in the initial test suite produced during initial development is inefficient, costly, and unacceptable in the maintenance phase [6]. Moreover, it does not consider obsolete and new test cases. In addition, it is often impractical due to the development cost and delivery schedule constraints [7]. An alternative approach, known as a selective retest strategy, assumes that not all parts of a software system are affected by changes [8]. Regression test selection in this case consists in selecting and running, from an initial test suite, a reduced subset of test cases in order to verify the behavior of modified
%U http://www.hindawi.com/journals/isrn.software.engineering/2013/420394/