Hi Graham,
One approach is to define your test content in an abstract base class. The test content would refer to an abstract property or method in the base class that is implemented by each subclass. You would need 100+ subclasses (one for each of the scripts that you want to test). Andy Campbell gives a concrete example of this approach here. This approach has the obvious disadvantage of requiring the definition of over 100 test classes; however, they would likely be quite simple. One benefit is that you can easily run the unit test for just one script file by itself. Alternatively, you could consider writing Test methods that operate on all the scripts using a for loop. In order to provide defect localization in the event of a test failure, you could construct a diagnostic that identifies the script being tested. Here's a rough outline of an example:
classdef exampleTest < matlab.unittest.TestCase
methods (Test)
function test1(testCase)
for i = 1:100
testCase.verifyThat(result_of_script_i, <constraint>, ...
sprintf('Testing script %d', i));
end
end
end
end
With this approach, you lose the ability to easily run the test for just one of the scripts by itself, but the test is fairly easy to write.
Hope this helps,
David
댓글 수: 0
댓글을 달려면 로그인하십시오.