Testing, Minimization, Transparency and Performance of System Test Case Prioritization: a Case Study of Regression Testing: Improving Regression Testing Performance with Test Case Prioritization Techniques

Abstract

In a software testing domain, different techniques and approaches areused to support the process of regression testing in an effective way. The mainapproaches are test suite minimization, test case prioritization and test caseselection. Test case prioritization techniques improve the performance ofregression testing, and arrange test cases in order to obtain maximum availablefault that is going to be detected in a shorter time. User-sessions and cookies are unique features of web applications thatare useful in regression testing because they have precious information aboutthe application state before and after any change that is made to the softwarecode. The main challenge is the effectiveness of average percentage faultdetection rate and time constraint in the existing techniques. Thus, in thisresearch the priority is given to clustering test cases that are performedbased on some criteria related to http requests which collected from thedatabase of server side. To verify the new technique some fault will be seededin subject application then applying the prioritization criteria on test casesfor comparing the effectiveness of average percentage fault detection rate andtime with existing techniques. Prioritization of test cases is generally done to reduce the cost ofregression testing. We prioritize our test cases so that those which are moreimportant, by some measure, are made to run earlier in the testing phase. Thereexists a large variety of prioritization techniques in the literature, we havebasically used coverage-based prioritization techniques (i.e., prioritizationin terms of the number of statements, path coverage, branch coverage and faultcoverage) controlling the field. A prioritized test suite which covers more than one coverage criteriais considered to be a stronger than those which cover only single coverage. Theproposed method was empirically studied for bank application and the resultsshow that the proposed work is more effective than the existing method.