C9: Unit Testing in C

Question 1

What are mistakes in software called?

Answer 1

bugs

Question 2

What is software verification?

Answer 2

Verification is making sure software fulfills its specification.

Question 3

What is software validation?

Answer 3

Validation means making sure the software fulfills the users’ needs (acceptance tests).

Question 4

What are the different methods of software verification?

Answer 4

Formal proofs, reviews, analysis and tests.

1. formal proofs…
2. reviews - code reviews, design reviews,…
3. analysis - static code analysis, dynamic resource usage measurements,…
4. Tests - model driven, unit tests, …

Question 5

What is test automation?

Answer 5

using a tool to execute
tests automatically (instead of manually)

Question 6

What is model-driven (automated) testing?

Answer 6

model-driven (automated) testing means using a (formal) model of the software under test to create test cases (and pass criteria) automatically.

Question 7

What is regression testing?

Answer 7

Regression testing is repeating existing tests again to make sure that modifications in previously tested software did not cause any new bugs or otherwise break the system.

Question 8

What are the different kind of tests?

Answer 8

1. Unit tests (test one unit), 2. integreation tests (test multiple units) and
2. system tests (test whole system)

Question 9

What are the different kinds of system tests?

Answer 9

1. functional tests
2. performance tests
3. stress tests
4. user interface tests
5. recovery tests
6. failover tests

Question 10

What are performance tests?

Answer 10

A type of system test: Performance test: tests behaviour of system (e.g. reaction time) in specified workload boundaries

Question 11

What are stress tests?

Answer 11

A type of system test: Stress test: performance test that overloads the system.

Question 12

What are recovery tests?

Answer 12

A type of system test: Recovery test: tests if system can recover
after (full) system failure.

Question 13

What are failover tests?

Answer 13

A type of system test: Failover test: tests if system works as
specified during (partial) system failure

Question 14

What is an integration test?

Answer 14

Integration testing checks the interplay of several units and checks whether they work together correctly.

Question 15

What is a resource test?

Answer 15

It is a test to see whether modules misbehave due to scarce resources (integration testing).

Question 16

What is unit testing?

Answer 16

Unit testing makes sure each unit works correctly, given the right input and assuming all other units are correct. Unit tests are also WHITE BOX tests to make sure that one individual unit works correctly independently of all other units.

Question 17

What is white box and black box testing?

Answer 17

white box testing- we have source code.

black box testing - we do not have source code.

Question 18

What are the F.I.R.S.T properties of unit tests?

Answer 18

Fast, Isolated, Repeatable, Self-Verifying, Timely:
Fast - measure in fraction of a second,
Isolated - single reason to fail,
Repeatable; obtain same results every time,
Self-Verifying; test fails or passes unambiguously,
Timely small time gab between writing test and code.

Question 19

What is a test double?

Answer 19

A test double impersonates some function, data, module or library during a test.  Other variations of a test double are fake, mock, stub.... For example:
test dummy
test stub
test spy
mock object
fake object
exploding fake

Question 20

When should you use test doubles?

Answer 20

 Hardware independence
 Inputs that are difficult to produce
 Slow collaborator
 Dependency on something volatile
 Dependency on something under
development
 Dependency on something that is difficult
to set up and configure

Question 21

How do we get rid of dependencies to other units or external factors (OS, HW)?

Answer 21

By isolating the Unit using test doubles instead of collaborators (some function, data, module or device outside the code under test that the code depends on.) This is possible by using

1. link-time substitution,
2. function/pointer substitution/dependency injection,
3. preprocessor substitution.

Question 22

What is link-time subsitution?

Answer 22

Link-time substitution is used to break dependency of a code under testing and its DOC/collaborator (i.e. function, etc. outside the code that the tested code depends on.) . It does this by:

1. providing different DOC implementation to linker when testing,
2. replacing DOC for whole unit test executable

Question 23

What is preprocessor substitution?

Answer 23

is used to break dependency of a code under testing and its DOC/collaborator (i.e. function, etc. outside the code that the tested code depends on.) . It does this by: It is a last resort if link-time sub or pointer sub doesn’t work.
use preprocessor directives in CUT source code to redefine names and/or include different header files when testing.

Question 24

How do you test functions that print a string?

Answer 24

You can recreate the output by reproducing the string and then using the function TEST_ASSERT_EQUAL_STRING()

Question 25

What is a mock object?

Answer 25

A mock object is used in unit testing to check whether indirect side effects of a function are correct.