Search Based Software Testing Hands-on

Randoop

1. Compile Foo.java
2. Invoke Randoop on class Foo to generate regression and error test cases (use the option to consider uncaught exceptions as errors)
3. Try to create another Java class Bar that contains an exception-raising branch that is very difficult for Randoop to execute

EvoSuite

1. Compile Foo.java and Stack.java
2. Use EvoSuite to generate branch-adequate JUnit test suites
3. Inspect the generated test cases

Local Search

1. The given skeleton will generate a random Python list of length LEN, each element in the random range(MAX), called hidden.
2. A Euclidean distance based fitness function is provided.
3. Write a hillclimbing algorithm that uses the fitness function to retrieve contents of the hidden list, along with the number of spent fitness evaluation.
4. Write the AVM algorithm, and compare the number of fitness evaluation spent.

Instrumentation for SBST

1. The template contains the example branches that we saw in the slides.
2. Rewrite the function foo so that, after each execution, the function updates appr and bdist with approach level and branch distance w.r.t. the target branch (HINT: you can add support functions that implements the branch distance recipe)

Combining Everything

1. Take the hill climbing algorithm from the Local Search task
2. Apply it to the manually instrumented function foo
3. See if we can generate the test data automatically
4. Consider what needs to be done if you want to do the instrumentation automatically (using Python ast module)