International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 03 Issue: 09 | Sept-2016 www.irjet.net p-ISSN: 2395-0072 © 2016, IRJET | Impact Factor value: 4.45 | ISO 9001:2008 Certified Journal | Page 517 An Empirical based Object Oriented Coverage Analysis Using XML Sasanko Sekhar Gantayat, Srinivas Prasad, K. Koteswar Rao Department of Computer Science & Engineering, GMR Institute of Technology, Rajam, Andhra Pradesh, India. sgantayat67@gmail.com, 1 srinivas.prasad@gmrit.org, koteswarrao.k@gmrit.org Abstract: Testing of Object-oriented software has a number of features that make it different from conventional software testing. With the increase in size and complexity of modern software products, the importance of testing is rapidly growing. In this paper, a new methodology is proposed to evaluate the code coverage, its effectiveness and compared its advantages over other traditional techniques. Keywords: Test Coverage, Object Oriented Coverage, Coverage Measure, Test Coverage Tools, XML. 1. Introduction to Test Coverage Test Coverage is the process of identifying the extent to which different test suites are appropriate in accessing a full complement of the source code in testing various components of the system. To the management and deployment team, it provides assurance on the comprehensive nature of the tests and provides insight into the areas where testing is inadequate, or where the labor expended on releasing a product can be reduced due to ineffective testing. It would be worthwhile to investigate the possibility of testing based on coverage analysis approach and study its effectiveness. It is the process which provides a measure of how well the test suite actually tests the software. The major aspect of coverage analysis are[6]:  Finding areas of a program not exercised by a set of test suites  Creating additional test cases to increase the coverage  Determining a quantitative measure of coverage which is an indirect measure of quality  Identifying redundant test cases that do not increase coverage 2. Basic Coverage Measures There are many coverage measures. Brief description about some of the basic coverage measures are given below [1,3,4,5,6,7,16,18]: Statement Coverage: This measure reports the uncovered statements as well as a percentage of statements that are covered. It might not detect the control faults. It tells whether the nodes in a control flow graph (CFG) are executed or not. It can not check whether all the branches are executed or not. CFG: It is a graphical representation of a program, in which each node consists of a set of statements that can be executed sequentially and edges are labeled with conditional statements. Branch Coverage: It explores whether all branches are executed or not. It checks for the edges of the CFG i.e., it checks for all the branches that are formed with if statement, for statement, while, do while, switch statement and exception handlers. It is also known as Decision coverage. It reports whether Boolean expressions tested in control structures evaluated to both true and false. Decision coverage: It reports whether Boolean expressions tested in control structures (like if statement, while statement) are evaluated to both true and false. The entire Boolean expression is treated as one true or false predicate although it contains logical - and or logical-or operators. In addition, it includes switch-statement cases, exception handlers and interrupts handlers. Condition Coverage: It reports the true or false outcome of each Boolean subexpression, separated by logical – operators if any. It measures the subexpressions independently of each other. This measure is similar to Decision coverage but has better sensitivity to the control flow. Path Coverage: It reports whether each of the possible paths is executed or not. A path is a unique sequence of branches from the function entry to the exit. It is also known as Predicate coverage. But due to the loop structures, many variations of this measure exist. It has two major disadvantages – the first one is the number of paths increases exponentially to the number of branches, the second one is that many paths are not possible to exercise due to relationships of data. Method Coverage: It gives information that each method is invoked or not. In Java there are some methods declared as abstract i.e., they do not have a body. Therefore, this has to be taken care by the test coverage analyzer.