Research On Automatically Generating Methods Of COM Testing Cases And Testing Scripts

Nowadays, component and Component-Based Software Development (CBSD) are gaining substantial interests in the software engineering community. A component-based approach can simplify the design and maintenance of large software systems, and also can increase reusing opportunities and productivity. However, components with low-quality or wrong configuration which used in large software systems will bring in many problems, even lead to disasters.Components used to be supplied by a third partner, so the source codes of them are normally not accessible. Then the traditional testing methods can't be used in it directly. Meanwhile the features of encapsulation and state are also increase the difficulties.As the kernel of testing, testing cases and scripts generating methods are always the emphases and hotspots in this area. Testing of component and CBS should take the most of the traditional testing methods as well as consider the features of component itself. Because the trait of state between component and object are alike, object-oriented testing methods can be used in testing component and CBS via some extend.Model is an object's constraints on state and messages which can communicate. Components are divided into nonmodal component and modal component via the introduction of model in component, and then it can be tested according to its own features. Based on the modal-feature of component, generating methods from testing cases to testing scripts are introduced respectively. For nonmodal components, boundary values and pair-wise coverage methods are used to generate testing cases automatically, with XML and reflection mechanism; for modal components, extended adjacency table and extended constraint table are defined, combining with stream-based testing methods, which are using to generate testing sequences, besides the methods used in nonmodal components. Testing Cases of Component Generating System (TCsCGS) is introduced based on the theory discussed above, validating the efficiency of the testing methods.