Reading Notes-Composition of software-> Testing

A running software consists of four parts

1. executableCode

2. Configuration

3. Runtime Environment

4. Data

Any part of the changes will change the behavior of the software.

In many teams, they often only test the code, but ignore it, or do not test the code completely. Therefore, many problems will be encountered during subsequent deployment.

 

Frequent small-scale tests (usually one test is triggered every time a check-in operation) can ensure the quality of the Code.

Here we need AAT (automatic acceptance test) and a process of continuous integration, construction, and testing.

Combined with appropriate manual testing, we can better ensure the quality of the software.

Our tests generally cover the following steps:

1. Successful Compilation

2. Successful Unit Tests

3. Meet software quality standards, such as code coverage, or other technical metrics

4. Pass functional tests, mainly including business-related functions

5. Pass non-functional tests, including performance and security features

6. Through exploratory testing, we hope to discover more problems and provide feedback and improvement.

Note: All tests should be as close as possible to the production stage, includingArticleAt the beginning

The Application of automated testing generally includes 1-5. If possible, it should be as automated as possible (whenever you start to do repetitive things, you should think about how to optimize it)

 

The second step is to perform an automated test on check-in each time, which reduces the version gap between the two tests. Only a small part of the changes will help you locate the problem accurately.

The automated testing report will be automatically sent to your mailbox shortly after you check in the code. You can quickly view the test feedback and decide the next step (fix the bug or continue ?)

In general, we force users to respond to certain feedbacks. For example, compilation failure is a big problem, he will prevent others from developing and automatically testing (for example, he cannot compile the code after the user obtains the latest code)