- Problem:
- There is a need to specify sets of test cases that together
achieve some specific testing requirement. At times these test cases must
be executed in a specific sequence in order to achieve that goal.
- Context:
- An individual test case presents a single stimulus to the
Object Under Test (OUT). A particular test case may operate properly but
leave the
OUT in a state in which further stimuli will not be handled properly.
An object can only be considered correct when it has been tested over its
complete life cycle. That is, after an object has received a set of messages
that has thoroughly exercised all of its intended actions, often denoted as all
paths through the state model of the object.
- Forces:
-
- automation of test cases- Application of test cases can be
partially automated by grouping in a single test method several test
case messages.
- interaction of messages- Method interaction can be tested by
execution of multiple
test cases on a single OUT rather than beginning each new test with a fresh
object. These actions could be programmed as a single method rather than
calling the other, existing, methods.
- Solution:
- Create test methods that serve as scripts for invoking
multiple test cases.
The public interface of each test class contains only test scripts
that the tester may invoke. The test case methods are protected from direct
access.
Each test script creates an OUT and so any test cases executed within
the test script are stimuli to that single OUT.
Some standard scripts that are created are the baseline test script, Create
a baseline test script, and functional, structural and interaction test scripts
that ensure that the OUT is tested to specified coverage criteria. For projects
with specific performance criteria, there would be a performance suite that
concentrates on defining specific types of loads on the system.
- Structure:
- A test script is simply a method, but one that has
responsibility for creating the OUT, applying the test cases and then
deleting the object under test.
void SimpTest::test_script9()
{
OUT = new CUT;
*ostr << "Start an interaction of test cases 4 and 6" << endl;
if (test_case4() && test_case6())
*ostr << "Interaction test case completed successfully" << endl;
else
*ostr << "Interaction test case completed unsuccessfully" << endl;
delete OUT;
}
- Participants:
- The CUT and the associated test class are the primary
participants. Methods in the test class either present a single stimulus to
the OUT or multiple stimuli.
- Applicability:
- Use the sequence test cases by using test scripts
pattern for tests that can be composed from other, simpler, tests or
when a test is investigating the effects of multiple messages on a single OUT.
- Benefits:
-
- Test scripts provide a means of using multiple test cases to build and test more
complex interactions.
- Various types of standard test coverages such as functional and
structural, required for the HIT analysis,
can be achieved by aggregating test cases by use of test scripts.
- Risks:
- Incorrect sequencing of test cases in scripts can result in
misinterpreted test results.
- Resulting Context:
- This compositional organization of test cases results
in an increased ease of automation of executing tests.
- Patterns Referenced:
- Pattern Test a method in the Context of a class
|