Evaluation of a new structural testing model for concurrent programs

• Grant number: 12/01901-9
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Effective date (Start): April 01, 2012
• Effective date (End): March 31, 2013
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Paulo Sérgio Lopes de Souza
• Grantee: Murilo Giacometti Rocha
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

Testing concurrent programs is inherently more difficult than testing sequential ones. Features such as communication and synchronization among processes, and nondeterminism impose problems in applying the traditional sequential testing criteria. However, some of these issues have been addressed with positive results. In particular, the TestPar project has been developed at ICMC/USP, in partnership with UFPR and the University of Southhampton, to adapt traditional criteria for concurrent programs in general. New criteria have been successfully applied for different languages and libraries as well as paradigms, with the support of the ValiPar software-testing tool. Nowadays, ValiPar is instantiated for PVM, MPI, ValiPThread, and BPEL. Still, these defined criteria considered only restrictive semantics of concurrency primitives, imposing some changes from one context to another. Thus, an extended test model was proposed recently supporting a wider variety of languages, libraries, and paradigms. This project is inserted in this context and aims to design a new ValiPar version for the generation and evaluation of elements to be covered by test cases, considering the extended model. Its focus relies on Vali_Elem and Vali_Eval modules mainly. It also intends to evaluate the efficiency of the test criteria, test execution, and the generality of the test model under a different context (Java language), by using this new version of the tool. In this manner, some of the current open issues of the TestPar project will be investigated.(AU)