Strategies for regression testing

Regression testing is a test activity that runs through all phases of the test. Its purpose is to verify that the defects that have been detected have not been correctly modified and that the process of modification has not caused new defects. The software will verify the regression test after the defects found in the test or other activities have been modified. You can use different strategies when doing regression testing. 2. What are the strategies?

(1) You can choose to repeat the test completely. All the test cases are fully executed on one side to confirm the correctness of the problem modification and whether the perimeter is affected by the modification.

disadvantage: Because the use case is executed all, it will increase the project cost, also affect the project progress. So it's hard to do it completely, so it leads to the regression test strategy

(2) The test can be selectively repeated. You can select a part to perform to confirm the correctness of the problem modification and whether the perimeter is affected after the modification.

So how do we choose use cases? Here are three ways to do this:

(1) Overwrite the modification method, and select all the use cases of the module to test the error module. This can only verify that the module is still defective, but there is no guarantee that the surrounding modules associated with it will not cause defects due to this change.

(2) The peripheral impact method, in addition to the error module of the use case execution, the surrounding and its associated modules of the use case also executed once, to ensure the quality of the regression test. Of course, we can also use a quantitative perspective to analyze the quality of the module, such as: After a series of regression tests above, to see if the residual defect rate is already within the allowable range, then we can end this regression test as a standard.

(3) The indicator reached the law, after the test is complete, to see if we have reached the established targets.

3. The process of regression testing

1. Develop regression testing strategies during the Test strategy development phase

2. Determine the regression test version

3. Regression test release, performing regression tests according to regression test strategy

4. Regression test pass, close defect tracking order

5. The regression test does not pass, and the defect ticket returns to the developer. Re-modify and do regression testing again.

Whenever a new module is added as part of the integration test, the software changes. New data flow paths are set up, new I/O operations may also occur, and new control logic may be activated. These changes may cause errors in functions that are normally working properly. In an environment that integrates test strategies, regression testing is a way to re-perform certain subsets of some already-tested tests to ensure that these changes do not propagate unforeseen side effects.

In a wider environment, successful test results (of any kind) are found to be errors, and errors are to be modified, and some aspects of the software configuration (Programs, documents, or data) are modified whenever the software is modified. Regression testing is used to ensure that changes (for testing or other reasons) do not lead to unpredictable behavior or other errors.

Regression testing can be done manually by re-executing a subset of all the test cases, or by using an automated capture playback tool. The Capture playback tool enables software engineers to capture test cases and then play back and compare them. The regression test set (a subset of the tests to be performed) includes three different types of test cases:

* Representative test cases capable of testing all the functions of the software.

* Additional testing specifically for software features that may be affected by the modification.

* Testing for modified software components.

In the process of integration testing, regression testing can become very large. Therefore, regression testing should be designed to test only the main functionality of the module in which the error occurred, and every time a modification is made, it is impractical and inefficient to re-execute all tests on each of the program's functions.

4. Finally, I would like to say

In fact, I've never done anything like this before, and I've been sharing what I've learned every week as a technical blog. Actually at the beginning really feel very bored, also feel no big use, but later my idea completely changed. This method not only can supervise our study progress, help us learn more about the content of software testing, but also can help students and netizens to exchange experiences with each other to share their daily learning knowledge. There is no problem as long as the search for a look in the keywords can see a lot of relevant professional friends share knowledge, which I feel that the benefits.

The next session of the study younger sisters ~ ~ When you do this homework welcome here to exchange experience yo ~ so ~ First Goodbye ~

Strategies for regression testing