MutationTestingTools

(also see CodeCoverageTools)

• Proteum [1993]: http://www.serc.net/report/tr168p.pdf The first mutation testing tool that implements all mutation operators designed for the ANSI C programming language http://www.serc.net/report/tr41p.pdf. The tool is available by writing to Professor Jose Maldonado [email protected]?.

• µJava (muJava) is a mutation system for Java programs. It automatically generates mutants for both traditional mutation testing and class-level mutation testing. µJava can test individual classes and packages of multiple classes. Tests are supplied by the users as sequences of method calls to the classes under test encapsulated in methods in separate classes. - http://www.ise.gmu.edu/~offutt/mujava/

• SQLMutation is a mutation system for database queries. It generates mutants of SQL SELECT statements from a Web browser or from other applications that consume a Web Service. - http://in2test.lsi.uniovi.es/sqlmutation/

• JesTer (for Java code and JUnit tests) (free, open source) ("strong" mutation testing) - http://jester.sourceforge.net

• PesTer (for Python code and PyUnit tests) (free, open source) ("strong" mutation testing) - http://jester.sourceforge.net

• Nester (for C# code and NUnit tests) (free, open source) ("strong" mutation testing) - http://nester.sourceforge.net

• insure++ (for C++ code) (commercial product) - http://www.parasoft.com/products/insure/papers/tech_mut.htm

• GCT does mutation coverage for C ("weak" mutation testing). (open source, standard distribution hasn't been ported to modern versions of Unix or Windows) - http://www.testing.com/tools.html

• One of the oldest Mothra tool is available - http://ise.gmu.edu/~offutt/rsrch/mut.html

• Plextest C++ ("strong" mutation testing, commercial product) - http://www.omnidigm.com/plextest/

• Heckle (for Ruby) (free, open source) - http://rubyforge.org/projects/seattlerb

MutationTestingTools

Jeff Offutt's answer: Mutation changes the program and then asks the tester to find a test that causes the mutant to fail (produce incorrect results). Strong mutation looks at final output to decide whether to kill the mutant, and weak mutation looks at the state of the program immediately after the mutated statement. Weak mutation is approximately 50% cheaper than strong. Results differ when we have "masking", that is, when an incorrect state does not result in incorrect output. This seems to be fairly rare.