As a programmer, I sometimes pursue the implementation and performance of functions, and pay little attention to program fault tolerance. At least I have considered this very little, although I will handle it wherever I know there will be problems, however, for normal tests, we are also verifying that our ideas are correct and can be delivered without any problems. However, when we actually deliver the tests, we find that, there are still differences between testers and programmers.
Recently, I made a small tool mainly to check the correctness of the company's documents and to provide the correct or not detection results. It can be said that the functions are not complex and the implementation process is relatively smooth (mainly involved in the collaboration of various departments, therefore, it is still a little time-consuming), the results are also good, and there is no problem with the individual test, so the customer test is delivered.
Result: The test result shows that an exception is not eaten. The exception is caused by the failure of the user system (in general, the dependent service programs are opened, but are forcibly closed during the test) only when the test results were received, there was only one idea: Program fault tolerance. In the past, tests were conducted under normal circumstances, but now in abnormal circumstances, the results will expose some problems, which may be the importance of testers. They will consider various scenarios and test cases to jointly safeguard various aspects of the program.