Design of a Selective Continuous Test Runner

Abstract

This study presents the design of a selective continuous test runner, which has not been done before. Previous studies present only a continuous test runner with random or semi-random test cases selection techniques. Here we present the concept of influence graph which is constructed directly by using source code and then use this influence graph to detect any influence on any test case and run the tests on the background. For that purpose three algorithms were designed; one for building the influence graph for the first time, another for enhancing the influence graph according to code modifications, and the third for marking relevant test cases for retesting. We created an Integrated Development Environment (IDE) for test purpose. The mentioned algorithms were implemented on this IDE. Experimental results show: (1) Influence graphs helped efficiently in detecting the changed test cases, (2) The proposed technique worked well with large projects, and (3) The selective continuous test runner helped in detecting logic deviations in a more effective and fast way than the regular test running schemas.