Automatic Test Scenario Generation for Object-Oriented and Web-Based Applications

Abstract

Software Testing is a process to explore different errors, bugs and missing requirements in the software. Before testing a software we need to identify the test scenarios. This thesis describes software testing and automatic test scenario generation using Unified Modeling Language ( UML) state machine diagram as a first contribution and automatic generation of test scenarios from server side source code as a second contribution. In the first contribution, we have developed an approach to generate test scenarios automatically from state machine diagram. We have used the state machine diagram as input and then we have produced the control flow graph of the system and based on that by using the Depth First Search (DFS) algorithm we have generated the test scenarios and the number of independent paths for the program. The generation of test scenarios is very important for test case generation of the system. We have used this approach during the design phase of the software development prior to the actual testing phase to reduce the cost and time of testing and later it will also require less effort for testing of the system.
In the second contribution, we have used white box testing technique for automatic generation of the test scenarios. We have used the source code as input for the system. Next, we have parsed the source code through the parser. Then from the parse tree, we have generated control flow graph of the system. Finally from the CFG by using DFS algorithm, we have generated the test scenarios for the program. Both of our contributions are automatic and do not need any human intervention.