The Research Of Cross-browser Compatibility Testing For Web Application

With web applications widely used in various fields and the number of browser which they run on is growing fast and the version of them is changing so frequently that cross-browser incompatibility for web applications(Cross-Browser Incompatibility,XBI)is being a serious problem.Currently,the existing methods fot it has large proportion of manual work that it causes low coverage and low efficiency.This thesis proposes an automated cross-browser compatibility testing method called XBCT(Cross-Browser Compatibility Testing).It based on DOM analysis and element(Screen Element)analysis with computer vision.It includes following work.1.This thesis studies on a page probe,which can achieve DOM information and screenshot.It traverses the DOM tree with the breadth first search algorithm,collects all the node data of the DOM,and extracts the specified information from it to build a node information table after Web page.It snapshots the page and saves the picture automatically.It automatically generates test case set based on the node information table and the screenshots.2.This thesis studies on a cross-browser compatibility testing method called XBCT.A method called cross-browser incompatible HTML5 feature detection is used,which improves the test coverage.The test is runned automatically according to the test case set,which raises the efficiency.At the same time,it can figure out the element apparence incompatibility problem with DOM information and EMD(Earth's Mover's Distance).It ignores the dynamic elements so as to reduce the false positive rate and enhances the ability to detect true positive issues.3.This thesis evaluates the method on a set of real Web applications.The results show that it can archive higher accuracy and better performance compared with that of existing methods.