In web application testing existing model based web test generators derive test paths from a navigation model of the web application, completed with either manually or randomly generated inputs. Test paths extraction and input generation are handled separately, ignoring the fact that generating inputs for test paths is difficult or even impossible if such paths are infeasible. In this thesis, we propose three directions to mitigate the path infeasibility problem. The first direction uses a search based approach defining novel set of genetic operators that support the joint generation of test inputs and feasible test paths. Results show that such search based approach can achieve higher level of model coverage than existing approaches. Secondly, we propose a novel web test generation algorithm that pre-selects the most promising candidate test cases based on their diversity from previously generated tests. Results of our empirical evaluation show that promoting diversity is beneficial not only to a thorough exploration of the web application behaviours, but also to the feasibility of automatically generated test cases. Moreover, the diversity based approach achieves higher coverage of the navigation model significantly faster than crawling based and search based approaches. The third approach we propose uses a web crawler as a test generator. As such, the generated tests are concrete, hence their navigations among the web application states are feasible by construction. However, the crawling trace cannot be easily turned into a minimal test suite that achieves the same coverage due to test dependencies. Indeed, test dependencies are undesirable in the context of regression testing, preventing the adoption of testing optimization techniques that assume tests to be independent. In this thesis, we propose the first approach to detect test dependencies in a given web test suite by leveraging the information available both in the web test code and on the client side of the web application. Results of our empirical validation show that our approach can effectively and efficiently detect test dependencies and it enables dependency aware formulations of test parallelization and test minimization.

Test Generation and Dependency Analysis for Web Applications

BIAGIOLA, MATTEO
2020-01-15

Abstract

In web application testing existing model based web test generators derive test paths from a navigation model of the web application, completed with either manually or randomly generated inputs. Test paths extraction and input generation are handled separately, ignoring the fact that generating inputs for test paths is difficult or even impossible if such paths are infeasible. In this thesis, we propose three directions to mitigate the path infeasibility problem. The first direction uses a search based approach defining novel set of genetic operators that support the joint generation of test inputs and feasible test paths. Results show that such search based approach can achieve higher level of model coverage than existing approaches. Secondly, we propose a novel web test generation algorithm that pre-selects the most promising candidate test cases based on their diversity from previously generated tests. Results of our empirical evaluation show that promoting diversity is beneficial not only to a thorough exploration of the web application behaviours, but also to the feasibility of automatically generated test cases. Moreover, the diversity based approach achieves higher coverage of the navigation model significantly faster than crawling based and search based approaches. The third approach we propose uses a web crawler as a test generator. As such, the generated tests are concrete, hence their navigations among the web application states are feasible by construction. However, the crawling trace cannot be easily turned into a minimal test suite that achieves the same coverage due to test dependencies. Indeed, test dependencies are undesirable in the context of regression testing, preventing the adoption of testing optimization techniques that assume tests to be independent. In this thesis, we propose the first approach to detect test dependencies in a given web test suite by leveraging the information available both in the web test code and on the client side of the web application. Results of our empirical validation show that our approach can effectively and efficiently detect test dependencies and it enables dependency aware formulations of test parallelization and test minimization.
15-gen-2020
File in questo prodotto:
File Dimensione Formato  
phdunige_4305194.pdf

accesso aperto

Tipologia: Tesi di dottorato
Dimensione 3.39 MB
Formato Adobe PDF
3.39 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11567/991607
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact