/soda

Primary LanguageC++GNU Lesser General Public License v3.0LGPL-3.0

Welcome to SoDA.

The aim of the SoDA (Software Development Analysis framework) library, toolset and program repository is to provide researchers and practitioners a framework with which various code coverage-based analyses can be performed in a unified environment, and which can be applied to large programs and test suites efficiently, and without limitations on programming language, granularity and coverage criteria.

The SoDA library contains the necessary data structures to efficiently store a coverage matrix for multiple test cases, the results of your test suite execution of multiple revisions and the changes in your program during multiple revisions. These data structures can be used to implement prioritization or other dependence analysis related algorithms.

Visit our wiki page to get more information about the project.

The library is released under LGPLv3 license.

TAM

We have created a graphical user interface, called Test suite Analysis and Manipulation, to provide easy access to the features of SoDA library. You can find its repository at: https://github.com/sed-szeged/tam

What's next?

  • Java + JUnit integration
  • Create a tool or toolset that can help in creating coverage and result binaries of programs written in Java and tested with JUnit. Make it as automatic as possible.
  • Integrate the location of the code elements in the source code into the coverage binaries
  • To be able to connect the measurements with the source code.

Send us your improvement ideas or help us to implement these features. Open an issue to start a discussion or send your pull request.

Getting started

Dependencies

  • cmake 2.8.0 or higher
  • Boost 1.49 or higher
  • googletest (for unit tests)

Suported platforms

  • Linux
  • Mac OSX
  • Windows (with MinGW 32bit)

Compile

mkdir build #Create a build directory next to the cloned repository.
cd build
cmake ../soda/src  # Pass -Dtest=ON argument if you want to execute the unit tests.
make

Command line programs

The library contains several command line programs:

  • instrumentserver
  • Creates SoDA coverage matrix when you executing your instrumented code.
  • test-suite-selections
  • Runs one or more selection algorithm on the test suite an computes the precision, inclusiveness of the selected test cases according to the results of the test executions.
  • test-selection-statistics
  • Computes several statistics of the test suite like density of coverage matrix, average changes per revision, etc.
  • test-suite-metrics
  • Computes the FL and FD metric values of the test suite.
  • test-suite-score
  • Computes the fault detection score and fault localization score of a test suite based on actual faults.
  • test-suite-reduction
  • Reduces the given coverage data with specified reduction algorithms.
  • binaryDump
  • Reads and saves data from the binary formats in a human readable format.
  • rawDataReader
  • Creates SoDA coverage, results and changeset matrices from raw data sources.

Sample application

You can find a sample prioritization algorithm in the src/cl/SoDATools/sample directory. It demonstrates some of the features available in the SoDA library. The algorithm implements a general test case prioritization based on the coverage matrix. The test cases with higher coverage will be at the beginning of the prioritized list.

SoDA Repository

We provide a program repository with tests and prepared analysis scripts. The repository contains the SoDA binary data files of various programs and already made measurements, which can be used for research purposes.

Authors and Contributors

SoDA is created and maintained by the Department of Software Engineering, University of Szeged.

Contribution

If you would like to contribute to the project, create an issue at our GitHub page or send a pull request. Every contribution is welcome!

We are using git flow to manage our repository. The origin/develop is our main development branch. origin/master is always in production ready state.