What does test case stability mean?

There are two concepts in stability testing.

One is stability test, which corresponds to anomaly test, that is, the test of how the system reacts when an abnormal situation occurs. Including:

1 interactive test, interference, such as incoming calls, text messages, low power, etc. These are actually mentioned in the functional test in the previous chapter.

2 Test abnormality, network disconnection, power failure, server abnormality, etc.

2. Stability test refers to performance test and stress test.

1 benchmark performance test, by pressing the response speed of server ports and clients in different network environments.

2 big data test, in a specific environment, the client updates a large number of data and personnel list at one time.

There are other articles that mention performance testing, in order to evaluate the spatio-temporal characteristics of APP (really profound, spatio-temporal, another 4-dimensional, 5-dimensional? ), including:

1 limit test: under various boundary pressure conditions, such as battery, storage, network speed, etc. To verify whether the app can respond correctly.

-Install the application when the memory is full.

-Run the application phone and turn it off.

-Disconnect from the network while running the application.

The content of article 1 repeats these points.

2 Response ability test: Test whether various operations in the app meet the user's response time requirements.

-Response time for application installation and uninstallation

-Influence time of various functions and operations of APP

3 stress test: is the system resource occupation abnormal under repeated long-term operation?

-the-app repeatedly installs and uninstalls to see if the system resources are normal (do it a few times at will, normal person, who repeatedly installs and uninstalls)

-Other functions are repeatedly operated to see whether the system resources are normal (this should be the case).

4 Performance evaluation: Evaluate the usage of system resources in typical user application scenarios.

Here, what is a typical user application scenario?

5. baseline test should not be a benchmark performance test: benchmark test with competitive products, product evolution comparison test, etc. (meaningless).

Brief steps: adb equipment-understand the package name -ADB Shell monkey-P package name -v running times (the combination of multiple parameters forms different use cases to obtain maximum coverage)-analyze the monkey log when it crashes or does not respond.

General monkey command (can be used directly in the project):

Adb shell monkey-pcom.jiochat.jiochatapp-throttle100-ignore-ignore-security-exceptions-ignore-native-crashes-monitor-native-crashes-v-v/kloc-. d:\b.log

Reproduce the bug: the keyword ANR in the monkey log is abnormal, and the previous events, especially the seed value, should be exactly the same, such as the name of the monkey -p package -v seed 0 500.

Log analysis: check whether there are keywords such as crash, find the context, make a simple analysis, and send the locatable error information to the developer.

This tool is used for pressure testing. Developers combine the log printed by monkey with the log printed by the system to modify the problems in the test. Monkey is a tool attached to SDK, and all events are randomly generated without anyone's subjectivity.