The Research Of Job Schedule System For GUI Software Test Which Is Based On Jenkins

Agile development method is a provided effective manner to adopt user requirements exchanging.Automation test is an important phase of agile.For GUI Software Test required to execute a large-scale test cases in parallel mode.And automation test need to response quality for the Software in time limitation and source lake.So need a test schedule system provide a effective schedule to achieve requirements in production environment.This paper research on an existing task schedule system which is based on Jenkins server.The system is used to manage automation test cases Jobs execution order and the manner of controlling virtual machines on server.In one hand,in the agile development mode need to do automation test in a short term.In the other hand,according to the features of GUI software automation testing focus on whether UI Objects are covered.There already much research on how to create schedule for all test cases completed and can show software quality by passed and failed test case count.And these research also shown that different algorithms may more effective for some special issue.And different target of test require adopted target evaluation function for finding not bad schedule.The paper show a new evaluation function to make original task schedule algorithm more effective.So the paper approach a function as a way for evaluation which is focus on the UI Objects covered rate in the whole expected UI Objects should be executed.There is a demo project and data developed to analyze the function.Comparing of the data results can shown the reference of the GUI cover rate as target function.When expected 100% test cases completed,the efficiency increase 42% from original manner.And when expected 80% test cases completed,the paper apply Bayes classification algorithm to enhance the manner.Finally,the efficiency increase high to 72%.And it is also shown that different strategies have influence to the manner.In summary,it is shown that the GUI coverage used as the target function can generate a expected schedule plan and approach the requirements in smoke test and daily build test and regression test.It is evidence that the evaluation of GUI cove rate function is effective for quick feedback the results of GUI software automation test quality.