94 lines
2.9 KiB
Markdown
94 lines
2.9 KiB
Markdown
|
# 16\. Custom attributes
|
|||
|
|
|
|||
|
|
## Learning objectives
|
|||
|
|
|
|||
|
|
In this section we will look at some ways of implementing custom NUnit attributes, which can be used to alter test execution.
|
|||
|
|
|
|||
|
|
## Intro and motivation
|
|||
|
|
|
|||
|
|
A powerful part of NUnit is that it is very extendable. One of the ways it can be extended is through custom attributes. An example is attributes that implement the [IWrapTestMethod](https://docs.nunit.org/articles/nunit/extending-nunit/ICommandWrapper-Interface.html) interface. This interface has a method for wrapping a `TestCommand`, which implements a method for executing. Normally these test commands do something, call execute on their inner command and then maybe do something again after the inner command is completed.
|
|||
|
|
|
|||
|
|
In the following three classes an `IWrapTestMethod` interface is implemented and used in a test:
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
public class MyAttribute : NUnitAttribute, IWrapTestMethod
|
|||
|
|
{
|
|||
|
|
public TestCommand Wrap(TestCommand command)
|
|||
|
|
{
|
|||
|
|
return new MyCommand(command);
|
|||
|
|
}
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
public class MyCommand : TestCommand
|
|||
|
|
{
|
|||
|
|
private TestCommand innerCommand;
|
|||
|
|
|
|||
|
|
public MyCommand(TestCommand command) : base(command.Test)
|
|||
|
|
{
|
|||
|
|
innerCommand = command;
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
public override TestResult Execute(ITestExecutionContext context)
|
|||
|
|
{
|
|||
|
|
Debug.Log("Before");
|
|||
|
|
var result = innerCommand.Execute(context);
|
|||
|
|
Debug.Log("After");
|
|||
|
|
|
|||
|
|
return result;
|
|||
|
|
}
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
public class MyTests
|
|||
|
|
{
|
|||
|
|
[Test]
|
|||
|
|
[MyAttribute]
|
|||
|
|
public void Test1()
|
|||
|
|
{
|
|||
|
|
Debug.Log("The test");
|
|||
|
|
}
|
|||
|
|
}
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
When running `MyTests.Test1` the following output is printed:
|
|||
|
|
|
|||
|
|
Test1 (0,017s)
|
|||
|
|
\---
|
|||
|
|
Before
|
|||
|
|
The test
|
|||
|
|
After
|
|||
|
|
|
|||
|
|
Other interfaces that custom attributes can implement are `IWrapSetUpTearDown`, `IApplyToContext`, and `IApplyToTest`.
|
|||
|
|
|
|||
|
|
## Exercise
|
|||
|
|
|
|||
|
|
At Unity we have a goal that an action should never take longer than 500 ms. In the [sample](./welcome.md#import-samples) `16_CustomAttributes` there is a class called `MyClass`, which has two methods. Both methods are supposed to return true. However someone has made a regression so that one of the two methods takes a long time to run.
|
|||
|
|
|
|||
|
|
The task is to create a new custom attribute, which detects if the test takes longer than 500 ms to run. If that happens, it should fail the test with a descriptive message. Apply that to the two existing tests.
|
|||
|
|
|
|||
|
|
## Hints
|
|||
|
|
|
|||
|
|
* You can use the class `System.Diagnostics.Stopwatch` to time how many miliseconds have passed.
|
|||
|
|
|
|||
|
|
## Solution
|
|||
|
|
|
|||
|
|
A full solution for the exercise is availiable at `16_CustomAttributes_Solution`.
|
|||
|
|
|
|||
|
|
The core of the solution is the execute method in the test command implementation:
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
public override TestResult Execute(ITestExecutionContext context)
|
|||
|
|
{
|
|||
|
|
var stopWatch = new Stopwatch();
|
|||
|
|
stopWatch.Start();
|
|||
|
|
var result = innerCommand.Execute(context);
|
|||
|
|
stopWatch.Stop();
|
|||
|
|
|
|||
|
|
if (stopWatch.ElapsedMilliseconds > 500)
|
|||
|
|
{
|
|||
|
|
result.SetResult(ResultState.Failure, $"Test took {stopWatch.ElapsedMilliseconds} ms. That is longer than 500ms!");
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
return result;
|
|||
|
|
}
|
|||
|
|
```
|