Testing Android Apps Via Guided Gesture Event Generation

In recent years,mobile devices play an increasingly important role in our daily lives.A large number of mobile apps are produced,which greatly facilitate people's lives.Quality assurance for mobile apps becomes a hot issue and it should be addressed with a high priority.Different from traditional desktop apps(which are mainly oper-ated by keyboard and mouse),mobile apps are mostly driven by touch gestures whose interactions are natural to human beings.However,existing testing techniques can-not determine relevant gestures during app testing,thus generating gesture events for efficient and effective testing of such apps remains to be a challenge.To take gestures into consideration,there are mainly three strategies for existing techniques:(1)consider no other gestures besides the click gesture,(2)feed the apps under test with random gestures,(3)exhaustively enumerate all possible gestures.It is clear that strategy(1)and strategy(2)lead to incomplete test coverage,while strategy(3)suffers from efficiency issues.As a result,we need a new technique which can realize efficient and effective testing with gestures considered.Our work can be summarized as follows:1.We focused on the Android system,which is one of the most prevalent mobile platforms for smartphones and tablets.With a mini empirical study,we catego-rized how gestures are recognized in Android apps.2.Based on the study result,we designed a static analysis technique to obtain the gesture information:each UI component's potentially relevant gestures,so as to realize guided gesture event generation and reduce the amount of gesture events to be delivered in the automated testing.3.We designed an automated dynamic testing technique for Android apps.It is based on the depth-first search algorithm and it can support different strategies of gesture event generation so as to better show the testing effect.This automated dynamic testing technique solves several specific problems during app testing.4.Based on the above static analysis technique and the automated dynamic testing technique,we implemented our approach as a prototype by the name of GAT.Experimental results on five demo apps and nine real-world apps show that GAT is both efficient and effective in covering more code as well as detecting more gesture-related bugs.