Skip to content
Corina Pasareanu edited this page Jan 14, 2019 · 10 revisions

Symbolic PathFinder

Symbolic PathFinder (SPF) combines symbolic execution with model checking and constraint solving for test case generation. In this tool, programs are executed on symbolic inputs representing multiple concrete inputs. Values of variables are represented as numeric constraints, generated from analysis of the code structure. These constraints are then solved to generate test inputs guaranteed to reach that part of code. Essentially SPF performs symbolic execution for Java programs at the bytecode level. Symbolic PathFinder uses the analysis engine of the Ames JPF model checking tool (i.e. jpf-core).

Features

Symbolic PathFinder

  • Performs symbolic execution of Java bytecodes

  • Handles complex math constraints, data structures and arrays, multi-threading, pre-conditions, strings (on-going work)

  • Applies to (executable) models and code

  • Generates test vectors and test sequences that are guaranteed to achieve user-specified coverage (e.g. path, statement, branch, MC/DC coverage)

  • Measures coverage.

  • Generates JUnit tests, Antares simulation scripts, etc. (output can be easily customizable)

  • During test generation process, checks for errors

  • Is flexible, as it allows for easy encoding of different coverage criteria

  • Is integrated with simulation environment (on-going work)

Applications

Test input generation for Java container classes, NASA guidance navigation and control (GNC) software; script generation for testing execution engines. Symbolic PathFinder has been used at Fujitsu Labs for testing Web applications - see [http://www.fujitsu.com/global/news/pr/archives/month/2010/20100112-02.html Fujitsu press announcement]. See also [http://spinoff.nasa.gov/Spinoff2010/ct_3.html NASA’s post].

Other info

Repository

The old Mercurial repository is on http://babelfish.arc.nasa.gov/hg/jpf/jpf-symbc

Clone this wiki locally