HSP: a hybrid selection and prioritisation of regression test cases based on information retrieval and code coverage applied on an industrial case study

Abstract

The usual way to guarantee quality of software products is via testing. This dissertation presents a novel strategy for selection and prioritisation of Test Cases (TC) for Regression testing. In the lack of code artifacts from where to derive Test Plans, this work uses information conveyed by textual documents maintained by Industry, such as Change Requests. The proposed process is based on Information Retrieval techniques combined with indirect code coverage measures to select and prioritise TCs. The aim is to provide a high coverage Test Plan which would maximise the number of bugs found. This process was implemented as a prototype tool which was used in a case study with our industrial partner (Motorola Mobility). Experiment results revealed that the combined strategy provides better results than the use of information retrieval and code coverage separately. Yet, it is worth to mention that any of these automated options performed better than the original manual process deployed by our industrial partner to create test plans.