| 摘要 |
A solution is proposed for testing a software application. The test includes the execution of a series of test cases, each one involving the application of a predefined test input to the software application. The software application generates a corresponding output in response to this test input. A result of the test case is determined by comparing the actual output provided by the software application with an expected output thereof. The expected output of the test case is determined automatically. For this purpose, multiple auxiliary sources are exploited, such as other software applications different from the one under test. Each auxiliary source receives a corresponding input, derived from the test input, which is intended to cause the auxiliary source to provide the same expected output as the software application. The expected output is then estimated according to the actual outputs provided by the different auxiliary sources.
|
