Pester Testing Test Guide
Who this is for
Cmdlet behavior is validated using the Pester testing framework. The purpose of this document is to create a single standard to maximize unit test coverage while minimizing confusion on expectations. What follows is a working document intended to guide those writing Pester unit tests for PowerShell.
Unit testing is done not only to validate that the block of code works as expected, but also to assist the developer to know precisely where in the code to look; in some cases, seeing the source code may inspire better unit tests. In many cases, a unit test is the only documented specification. Fortunately, the MSDN is a great source of information about Cmdlets.
Test suites need to be created and many cmdlets added and unit-tested. The following list is to be used to guide the thought process of the developer in writing a suite in minimal time, while enhancing quality.
Test suites should proceed as functional and system tests of the cmdlets, and the code treated as a black box for the purpose of test suite design.
Portability
Some tests simply must be tied to certain platforms. Use Pester's
-Skip
directive on an It
statement to do this. For instance to run
the test only on Windows:
It "Should do something on Windows" -Skip:($IsLinux -Or $IsOSX) { ... }
Or only on Linux and OS X:
It "Should do something on Linux" -Skip:$IsWindows { ... }
Use of Mocks
It is often necessary for the code to interact with the system or other components. When possible, use Mock objects to facilitate this in order to minimize external dependencies. Note: creating a Mock in Powershell on Linux causes PowerShell to look at the Mock, never actually hitting any C# code. Cmdlets cannot be tested using Mocks.
Aliases
Each cmdlet with an alias must be tested with all of its aliases at least once to verify the code path calls the original function.
Testing Standards
Readability
Every effort should be made to maximize readability of code. Code is written for the developer in the future to debug- not for the developer writing the code.
- When assertions are on consecutive lines, the pipes should line up:
MyFirstCondition | Should Be 0
MySecondCondition | Should Be 1
This is less readable than:
MyFirstCondition | Should Be 0
MySecondCondition | Should Be 1
So the second section of code should instead be used. The same style should be followed for assignments of variables on consecutive lines:
$var1 = <expression 1>
$variable2 = <expression 2>
$var3 = <expression 3>
$typeCollection1 = <expression 4>
$object1 = <expression>
... etc
is much less readable than
$var1 = <expression 1>
$variable2 = <expression 2>
$var3 = <expression 3>
$typeCollection1 = <expression 4>
$object1 = <expression 5>
... etc
So all assignment statements must be aligned.
Other style standards are no less important to readability of the code:
-
Use readable and meaningful variable name when assigning variables.
-
Do not make large functions. Tests should be simple: define -> manipulate -> assert
-
Do not use tabs. Tabs are rendered differently depending upon the machine. This greatly affects readability.
-
Remove the first 3 auto-generated lines of each .Tests.ps1 file. This is created automatically by Pester and is unnecessary. Each .Test.ps1 file should begin with a Describe block.
-
Discard the auto-generated function file that is generated in tandem with the .Tests.ps1 file
-
Name the test file "Test- when you create a new test fixture.
-
Each test describes a behavior- use the word "Should" at the beginning of each test description- so it reads "It 'Should..."
Basic Unit Tests
The following table should suffice to inspire in the developer sufficient content to create a suite of tests.
test # | test name | entry criteria/setup | exit criteria/assertion |
---|---|---|---|
01 | Should be able to be called | without params (if applicable) | no throw |
02 | Should be able to be called | minimal required params | no throw, expected output |
03 | Should be able to use the X alias | minimal required params | no throw, expected output |
04 | Should return the proper data type | required params | no throw, proper data type |
05 | Should be able to accept piped input | piped input | expected output |
06 | Should be able to call using the X parameter | use X parameter | no throw, expected output |
07 | Should be able to call using the Y parameter | use Y parameter | no throw, expected output |
08 | Should be able to call using the Z parameter | use Z parameter | no throw, expected output |
09 | Should throw under condition X | create condition X | Throw error x |
10 | Should throw under condition Y | create condition Y | Throw error y |
11 | Should throw under condition Z | create condition Z | Throw error z |
These are the basic unit tests required to verify the functionality of any Cmdlet. If the above questions cannot be answered for each Cmdlet, then they cannot be verified to work.
Look at the existing suites of pester tests located within this directory and use that as inspiration.
Running Pester Tests
Go to the top level of the PowerShell repository and run:
Start-PSPester
inside a self-hosted copy of PowerShell.