Lab Assignment 3 Unit and Pairwise Testing (White-box Testing #1)

ECE 322 Lab Assignment 3
Unit and Pairwise Testing (White-box Testing #1)
Overview
The objective of this laboratory assignment is to become familiar with the rudimentary
techniques of white-box testing, specifically unit testing. Also, this assignment introduces us to
the pairwise test case generation tools. This lab makes use of Python, Pychram, the unittest
testing framework and allpairspy pairwise generation python package tool .
Introduction
White-box Testing:
White-box testing focuses on testing the internal structure and implementation details of the
application. In contrast to black-box testing, this style of testing requires the tester to have
detailed knowledge of the internal structure of the software under test, usually in the form of
direct access to the source code.
In this lab we will concentrate on control flow testing where we consider four fundamental
coverage criteria: statement coverage, branch coverage, condition coverage, and path coverage.

Pairwise Testing: Instead of considering all possible combinations of input variables, pairwise
testing provides an efficient way to construct a set of test cases that covers all combinations of
the test data for each pair of variables. The Orthogonal Array Test Strategy (OATS) is a
widely used technique of testing the pair-wise interactions of variables. An orthogonal array can
be expressed in the following way:
1 2 3
1 1 1 1
2 1 2 2
3 2 1 2
4 2 2 1
Frequently, the exact orthogonal array we need for the problem at hand is not available, in this
scenario we can select an orthogonal array which is slightly “too large” and modify it to suit our
needs, or use approximation of orthogonal arrays to generate reduced, though perhaps not
minimal, test cases. Due to the popularity and effectiveness of pairwise testing, many tools have
been developed to help generating pairwise test cases without the need for the proper orthogonal
array to be readily available, or the correct array to be time consumingly found in a database.
One such tool, is allpairspy python library which computationally calculates approximations to
orthogonal arrays which are then used to generate pairwise test cases.
The tool can be obtained at: https://pypi.org/project/allpairspy/
documentation is available at: https://pypi.org/project/allpairspy/, and
https://github.com/thombashi/allpairspy
A simple python script is available as an example on eclass
Note that the results produced from allpairspy are algorithmic approximations to orthogonal
arrays and may be less efficient in terms of the number of test cases than using a true orthogonal
array
Preparation: Make sure you have Pychram, Python3 and Virtualenv/Pipenv installed on your
account/computer. The code for this experiment is available on the class website and can be
imported into a Pychram python project once extracted. You may need to install some python
packages in order for the test cases to compile and run, check the source file to find all the required
python packages.
Task 1 (70marks)
The bisection method is a simple way to solve equations in the form of f(x) = 0 given an interval
in which the root can be found. The source for this project contains some code which executes
the bisection method to find the root of an equation. The source code is available on the class
website. First, draw a control flow graph for the algorithm, then generate test cases that comply
with:
● Statement coverage
● Branch coverage
● Execution each loop body at least twice (enter the loop and repeat)
Submit your test cases as unittest tests and compare results based on the above coverage criteria
in your written report. Constructors and error cases are included in the requirements for code
coverage. For your report you must include both:
● The test cases code which can be executed against the provided lab project
● A test case table similar to those used in black box testing containing a meaningful
description for each test case, the expected result, and the actual result
● The HTML generated coverage reports
In addition to the requirement of code coverage, keep in mind that for full marks your test cases
are required to assert meaningful things regarding the functionality of the code and the
results of your tests. For example, a successful use of the bisection method should assert that a
root value is returned from the algorithm within the given error threshold. Provide a brief
discussion on the effectiveness of these coverage criteria. Based only on the coverage criteria,
how confident are you in the bug free nature of the code? Do tests fulfilling these criteria
adequately cover the full functionality of the algorithm? How many tests would be required to
fulfil path coverage?
As always, your test cases should be executed, and the results recorded into a test case table to
be handed in with your report. Failed test cases must be highlighted, and the reasons for failure
must be identified.
Task 2: (30 marks)
Assume we have a system with three independent variables (A, B, and C). Each variable has
three possible values (0, 1, 2).
Your task is to:
● Identify the set of test cases for this problem based on this tool, reduced if necessary
● Compare the use of orthogonal arrays to randomly generated combinations
And
● Create a valid input for the problem
● Use the tool to generate test cases
● Comment on the effectiveness of this type of tool in test case generation. Compare the
use of orthogonal arrays to the allpairspy estimation tool. There is no application under
test for this component, this is a conceptual exercise
Include the generated test cases in your lab report as well as your observations on this method.
As there is no application to test, test cases need not be executed. Comment on the effectiveness
of using these types of tools to generate pairwise test cases, and on the effectiveness of pairwise
testing in general. What types of errors are caught by pairwise testing, what types are missed?
Are the weaknesses of the method justified by the test case efficiency increase when compared to
exhaustive testing? How many test cases would be required to cover all combinations of input
values in the example application?
Lab report:
Must be typed, no handwritten versions will be accepted. Follow the general format as included
in the lab guideline. Submit all codes. Your report should include:
● Your write-up and any required diagrams
● The results of your test cases in table format
● The HTML generated coverage reports
● Any conclusions you have drawn from these test cases regarding the applications under
test.
● Your test case code as Python files to be easily executed against the provided source for
validation.
Clearly indicate all failed test cases and explain the cause of these failures.