| During the evolution of Web applications,due to changes in the structure and content of certain graphical user interface(GUI),some tests cannot be run directly,and should be repaired to be reused.Existing Web application test repair techniques mostly use the method of simply finding and matching elements with the same or similar attributes in the old page.Since the functional process to be tested and the semantic information of the page elements are not considered,there are problems such as difficulty in repairing and low repair rate.To address the above problems,through the fusion of the relevant semantic information,this thesis proposed an automated repair technique based on intelligent search.This technique first runs the test script on the old version of the application to extract element information of related elements in the event sequence,and generate a keywords sequence.Then,it semantically guides the state crawling using the keywords sequence to make it preferentially cover states related to the function and perform effective path repair.At the same time,the element semantic similarity is introduced in the similarity calculation of page elements in the new version to further improve the repair rate and the effectiveness of the repair.On the basis of the above work,this thesis implements an automated repair tool for Web application test scripts based on keywords-sequence-guided – ARTER.In order to verify the effectiveness of the technique in this thesis,a total of 66 versions of 15 open source applications were selected for experimental research.The results show that compared with other repair tools,WATER and VISTA,ARTER increased the repair rate of the broken test scripts by 24.8-33.9%,with an average repair rate of 86.8%.The contribution of path repair accounted for 16.9% of the total repair.Compared with the shortest path repair strategy,the semanticguided path repair strategy of ARTER greatly improved the effectiveness of path repair,with an average efficiency of 84.8%. |
PDF Full Text Request