These days there are also articles on the 51test that are viewed in some articles about the relationship between development and testing, very much.
Most of the points of view will describe the development and testing between the majority of problems, such as the development of the test love to find fault, like the dead, testing that development does not understand testing, quality awareness is not strong, and so on.
Today, in my work, I experienced one thing. The project is about to be released, the development of a module of a function of the implementation of the other logic, and this change as the test manager of this project I am transparent, I do not know that development has changed this logic. But fortunately during the project smoke testing phase, I wrote the automation scripts for all the modules ' main threads and carried out daily regression tests. In today's discovery of some of the script execution failed, troubleshooting reasons found that changed the logic of the local foot of the report is wrong, so and development confirmation, is the change logic.
Then I went through several other failed scripts and found that another text input restriction logic changed.
In this case, I think for development, the change of function is too casual, in their consciousness there is no complete quality consciousness, think that their change is only a small function, there is no need to know the test. However, it is often the development of this habit, but eventually lead to online failure.
For the development of this habit, there are several corresponding measures to monitor, one is to monitor each of their submitted code, two is through the automated script for continuous integration, constantly check the function of each module of the system. Of course, fortunately there is the latter, let me find out where this problem lies.
Here, although the issue was found before the release, but we think, if we do not have this condition, after several rounds have been tested, development and secretly changed the other functions, the tragedy will happen sooner or later.
Thinking over and over again, I feel like I'm still playing the black guy, throwing these questions out in the project Exchange group, and showing my point of view: to develop random change functions before they are released, and to test the quality of the project to ensure that the test is confident.
Development also throws its own point of view, said this is some small function, there is no need to tell the test everything, think I was in the dead. But I think it's not just a matter of functional size, but a sense, a habit. Even though I also feel that these small changes do not have a big impact on the system, but I am very afraid of the development of this habit, perhaps one day, because of this situation, people are not as good as the days, after the release because of the small function caused major problems, this situation is not impossible, I would rather be a pessimist.
For development, testing has the responsibility to eliminate the development of this habit, to cultivate their good quality awareness, now very popular defect prevention, I think, if the development also has a good quality awareness, is not in the effective prevention of defects.