SQL Data Generation to Enhance Search-Based System Testing

GALEOTTI, JUAN PABLO; ANDREA ARCURI

Praga

Conferencia; Genetic and Evolutionary Computation Conference, GECCO 2019; 2019

ACM

Automated system test generation for web/enterprise systems require either a sequence of actions on a GUI (e.g., clicking on HTMLlinks and form buttons), or direct HTTP calls when dealing withweb services (e.g., REST and SOAP). When doing white-box testingof such systems, their code can be analyzed, and the same typeof heuristics (e.g., the branch distance) used in search-based unittesting can be employed to improve performance. However, we-b/enterprise systems do often interact with a database. To obtainhigher coverage and find new faults, the state of the databases needsto be taken into account when generating white-box tests. In thispaper, we present a novel heuristic to enhance search-based soft-ware testing of web/enterprise systems, which takes into accountthe state of the accessed databases. Furthermore, we enable thegeneration of SQL data directly from the test cases. This is usefulfor when it is too difficult or time consuming to generate the rightsequence of events to put the database in the right state. And it isalso useful when dealing with databases that are read-only forthe system under test, and the actual data is generated by otherservices. We implemented our technique as an extension of theEvoMaster tool, where system tests are generated in the JUnitformat. Experiments on five different RESTful APIs show that ournovel technique improves code coverage significantly, even by upto +18%.