| 摘要 |
An automatic testing technique generates an incremental coverage report indicating coverage only for portions of the program that are unreached by previous testing-allowing software testing processes to concentrate upon finding tests which increase the test coverage and reducing the sheer bulk of the coverage output. A computer program to be tested is parsed and "instrument" code is inserted at appropriate parts of the program. The instrumented program is then typically executed multiple times with different input conditions in an attempt to "exercise" and test different parts of the program. For each execution, the instrument code inserted by the instrumentation process causes coverage results to be generated and stored. Such coverage results may include information about which parts of the program have been reached by the current execution that had not previously been reached by earlier executions. This "incremental" coverage results output is useful in specifying the value of a current test without flooding the user with volumes of information specifying all parts of the program reached by the current execution.
|
