Abstract:

In the software industry, nowadays automatic testing methods are widely used in GUI testing, which largely depends on manual work, and features low ad  aptability to software variation, and lacks systematic management as to testing cases and coverage. Accordingly, a more highly automated GUI testing sys tem which is based on event-flow graphs is proposed under the framework of Microsoft UI Automation, where the automation tool is first used to carry out  the reverse engineering for testing samples so as to obtain a majority of event-flow graphs and expected testing results. During the process of the formation of the testing cases, two strategies are adopted. For the daily smoke testing, an ant colony optimization algorithm is used to go through all theevent points in the event-flow graphs and establish a sequence as a testing case. For the deep regression testing, a spanning tree obtained by the goal-directed searching approach is explored to obtain the testing cases from the goal point to the outset point. And these cases are applied to test the new GUI. The results of experiments illaminate that the   improvement for the GUI testing automation in Windows systems.

Key words: automatic software testing, graphic user interlace, event-flow graph