Package edu.berkeley.cs.jqf.fuzz

Class FuzzRunner

java.lang.Object

edu.berkeley.cs.jqf.fuzz.FuzzRunner

public final class FuzzRunner
extends Object

The fuzzing loop, independent of any test framework.

For each input offered by the Guidance, the loop generates arguments through an ArgumentsGenerator, runs one trial through a TrialExecutor, classifies the outcome with a ResultClassifier, and reports it back to the guidance. None of these collaborators expose a test-framework type, so the loop carries no dependency on JUnit or any generator library.

When the campaign ends, the loop rethrows a single collected failure as-is, or wraps several in a MultipleFailuresError.

Constructor Summary

Constructors

Constructor	Description
FuzzRunner(ArgumentsGenerator argumentsGenerator, TrialExecutor trialExecutor, Guidance guidance, ResultClassifier classifier, List<Class<?>> expectedExceptions, Observability observability, boolean skipExceptionSwallow)	Creates a runner for one test method.

Method Summary

Modifier and Type	Method	Description
void	run()	Runs the fuzzing loop until the guidance stops offering input.
static FuzzResult	run(Class<?> testClass, String testMethod, Guidance guidance)	Runs a fuzzing campaign for a resolved test method.
static FuzzResult	run(String testClassName, String testMethod, Guidance guidance)	Runs a fuzzing campaign for a test method, loading the class with the system class loader.
static FuzzResult	run(String testClassName, String testMethod, ClassLoader loader, Guidance guidance)	Runs a fuzzing campaign for a test method, loading the class with the given class loader.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

FuzzRunner

public FuzzRunner(ArgumentsGenerator argumentsGenerator,
 TrialExecutor trialExecutor,
 Guidance guidance,
 ResultClassifier classifier,
 List<Class<?>> expectedExceptions,
 Observability observability,
 boolean skipExceptionSwallow)

Creates a runner for one test method.

Parameters:

argumentsGenerator - generates each trial's arguments

trialExecutor - runs a single trial with proper lifecycle

guidance - drives input selection and observes results

classifier - maps trial throwables onto engine outcomes

expectedExceptions - exception types declared by the test method; a thrown subtype counts as success

observability - collector for optional observability logging

skipExceptionSwallow - if true, declared exceptions count as failures rather than success

Method Details

run

public void run()
         throws Throwable

Runs the fuzzing loop until the guidance stops offering input.

Throws:

Throwable - a single trial failure, a MultipleFailuresError for several, or a GuidanceException if fuzzing was interrupted

run

public static FuzzResult run(String testClassName,
 String testMethod,
 Guidance guidance)
                      throws ClassNotFoundException

Runs a fuzzing campaign for a test method, loading the class with the system class loader.

Parameters:

testClassName - the fully qualified test class name

testMethod - the @Fuzz or @DiffFuzz method to run

guidance - the fuzzing guidance

Returns:

the campaign result

Throws:

ClassNotFoundException - if the test class cannot be loaded

run

public static FuzzResult run(String testClassName,
 String testMethod,
 ClassLoader loader,
 Guidance guidance)
                      throws ClassNotFoundException

Runs a fuzzing campaign for a test method, loading the class with the given class loader.

Parameters:

testClassName - the fully qualified test class name

testMethod - the @Fuzz or @DiffFuzz method to run

loader - the class loader for the test class (e.g. an instrumenting loader)

guidance - the fuzzing guidance

Returns:

the campaign result

Throws:

ClassNotFoundException - if the test class cannot be loaded

run

public static FuzzResult run(Class<?> testClass,
 String testMethod,
 Guidance guidance)

Runs a fuzzing campaign for a resolved test method.

The argument generator and the trial executor are discovered through ServiceLoader, so this entry point depends on no test framework. The guidance's instrumentation callback is registered before the entry point is traced, mirroring the JUnit 4 orchestrator.

Parameters:

testClass - the test class

testMethod - the @Fuzz or @DiffFuzz method to run

guidance - the fuzzing guidance

Returns:

the campaign result