Q-SFL

A qualitative-reasoning-based extension to SFL. This project provides a framework for partitioning method parameters and return types according to their runtime value as a way to mitigate ambiguity grouping and coincidental correctness. Based upon work published at IJCAI-ECAI'18.

Compilation and Installation

To compile the project and install the project in your local maven repository, simply run the command:

mvn install

Usage

Add the following to a Java project's pom.xml:

<build>
  <pluginManagement>
    <plugins>
      ...
      <plugin>
        <groupId>pt.up.fe.qsfl</groupId>
        <artifactId>qsfl-maven-plugin</artifactId>
        <version>0.1-SNAPSHOT</version>
        <configuration>
          <valueProbes>false</valueProbes>
        </configuration>
      </plugin>
      ...
    </plugins>
  </pluginManagement>
</build>

To run the project's test cases and perform the Q-SFL instrumentation, execute the command:

mvn qsfl:test

For an usage example, please refer to aperez/q-sfl-example.

Value Probes

If the valueProbes configuration parameter of Q-SFL is enabled, parameter values for function invocations will be recorded.

Caveats

Junit3 and Junit4 Compatibility

Although qsfl-maven-plugin works for both Junit3 and Junit4 test cases, please make sure that the Junit dependency included in the project's pom.xml is at least version 4.6.

This requirement is due to the fact that Junit only provides their org.junit.runner.notification.RunListener API after version 4.6. The listener API is used so that per-test coverage can be gathered. For Junit3 test cases, the appropriate test runner will still be used.

Currently, this plugin also does not support test execution forking.