Research On GUI Test Method Based On Multi-level Morphological Model

Model-based GUI testing(MBGT)approaches are efficient since their test cases can be generated automatically,but most MBGT models are fixed,that is,these models can only view the GUI in a fixed way.The multilevel morphology model(MMM)allows testers to explore the morphological differences of GUI model by enlarge the model context to an acceptable range,therefore,increase the fault detection effectiveness.However,MMM can only variates integrally while some parts of the GUI may need more attention.With the increasing level of MMM,the model become more and more complex and harder to process.Additionally,in model-based mutation testing,a large number of mutants are invalid.In this paper,we proposed a MMM segmentation approach which enables the MMM to variate partially and a relevant test case generation strategy.Based on event classification,this approach can segment and simplify the original model.According to the characteristics of the segmented model,using BFS and CPP(Chinese Postman Problem)solving algorithms to generate test cases.In mutation testing,treat the segmented model's boundary as constrains to mutation operation,eliminate all equivalent and invalid mutants.This approach segmented and simplified the GUI graphic model,allows testers to focus on the important parts of the model,and meanwhile,reduce the number and length of test cases,make MMM more agile and efficient.The result of the experiment indicated that the segmented MMM has almost the same fault detection effectiveness as its original model.As the model level increased,the segmented MMM will become more efficient than its original model.