dotnet test

Article
03/27/2024

This article applies to: ✔️ .NET Core 3.1 SDK and later versions

Name

dotnet test - .NET test driver used to execute unit tests.

Synopsis

dotnet test [<PROJECT> | <SOLUTION> | <DIRECTORY> | <DLL> | <EXE>]
    [--test-adapter-path <ADAPTER_PATH>]
    [-a|--arch <ARCHITECTURE>]
    [--artifacts-path <ARTIFACTS_DIR>]
    [--blame]
    [--blame-crash]
    [--blame-crash-dump-type <DUMP_TYPE>]
    [--blame-crash-collect-always]
    [--blame-hang]
    [--blame-hang-dump-type <DUMP_TYPE>]
    [--blame-hang-timeout <TIMESPAN>]
    [-c|--configuration <CONFIGURATION>]
    [--collect <DATA_COLLECTOR_NAME>]
    [-d|--diag <LOG_FILE>]
    [-f|--framework <FRAMEWORK>]
    [-e|--environment <NAME="VALUE">]
    [--filter <EXPRESSION>]
    [--interactive]
    [-l|--logger <LOGGER>]
    [--no-build]
    [--nologo]
    [--no-restore]
    [-o|--output <OUTPUT_DIRECTORY>]
    [--os <OS>]
    [--results-directory <RESULTS_DIR>]
    [-r|--runtime <RUNTIME_IDENTIFIER>]
    [-s|--settings <SETTINGS_FILE>]
    [-t|--list-tests]
    [-v|--verbosity <LEVEL>]
    [<args>...]
    [[--] <RunSettings arguments>]

dotnet test -h|--help

Description

The dotnet test command is used to execute unit tests in a given solution. The dotnet test command builds the solution and runs a test host application for each test project in the solution using VSTest. The test host executes tests in the given project using a test framework, for example: MSTest, NUnit, or xUnit, and reports the success or failure of each test. If all tests are successful, the test runner returns 0 as an exit code; otherwise if any test fails, it returns 1.

Note

dotnet test was originally designed to support only VSTest-based test projects. Recent versions of the test frameworks are adding support for Microsoft.Testing.Platform. This alternative test platform is more lightweight and faster than VSTest and supports dotnet test with different command line options. For more information, see Microsoft.Testing.Platform.

For multi-targeted projects, tests are run for each targeted framework. The test host and the unit test framework are packaged as NuGet packages and are restored as ordinary dependencies for the project. Starting with the .NET 9 SDK, these tests are run in parallel by default. To disable parallel execution, set the TestTfmsInParallel MSBuild property to false. For more information, see Run tests in parallel and the example command line later in this article.

Test projects specify the test runner using an ordinary <PackageReference> element, as seen in the following sample project file:

<Project Sdk="Microsoft.NET.Sdk">

  <PropertyGroup>
    <TargetFramework>net6.0</TargetFramework>
    <Nullable>enable</Nullable>
  </PropertyGroup>

  <ItemGroup>
    <PackageReference Include="Microsoft.NET.Test.Sdk" Version="17.10.0" />
    <PackageReference Include="xunit" Version="2.8.1" />
    <PackageReference Include="xunit.runner.visualstudio" Version="2.8.1" />
  </ItemGroup>

</Project>

Where Microsoft.NET.Test.Sdk is the test host, xunit is the test framework. And xunit.runner.visualstudio is a test adapter, which allows the xUnit framework to work with the test host.

Implicit restore

You don't have to run dotnet restore because it's run implicitly by all commands that require a restore to occur, such as dotnet new, dotnet build, dotnet run, dotnet test, dotnet publish, and dotnet pack. To disable implicit restore, use the --no-restore option.

The dotnet restore command is still useful in certain scenarios where explicitly restoring makes sense, such as continuous integration builds in Azure DevOps Services or in build systems that need to explicitly control when the restore occurs.

For information about how to manage NuGet feeds, see the dotnet restore documentation.

Workload manifest downloads

When you run this command, it initiates an asynchronous background download of advertising manifests for workloads. If the download is still running when this command finishes, the download is stopped. For more information, see Advertising manifests.

Arguments

PROJECT | SOLUTION | DIRECTORY | DLL | EXE
- Path to the test project.
- Path to the solution.
- Path to a directory that contains a project or a solution.
- Path to a test project .dll file.
- Path to a test project .exe file.
If not specified, the effect is the same as using the DIRECTORY argument to specify the current directory.

Options

Warning

Breaking changes in options:

Starting in .NET 7: switch -a to alias --arch instead of --test-adapter-path
Starting in .NET 7: switch -r to alias --runtime instead of --results-directory

Warning

When using Microsoft.Testing.Platform, please refer to dotnet test integration for the supported options. As a rule of thumbs, every option non-related to testing is supported while every testing-related option is not supported as-is.

--test-adapter-path <ADAPTER_PATH>

Path to a directory to be searched for additional test adapters. Only .dll files with suffix .TestAdapter.dll are inspected. If not specified, the directory of the test .dll is searched.

Short form -a available in .NET SDK versions earlier than 7.

--arch <ARCHITECTURE>

Specifies the target architecture. This is a shorthand syntax for setting the Runtime Identifier (RID), where the provided value is combined with the default RID. For example, on a win-x64 machine, specifying --arch x86 sets the RID to win-x86. If you use this option, don't use the -r|--runtime option. Available since .NET 6 Preview 7.

--artifacts-path <ARTIFACTS_DIR>

All build output files from the executed command will go in subfolders under the specified path, separated by project. For more information see Artifacts Output Layout. Available since .NET 8 SDK.

--blame

Runs the tests in blame mode. This option is helpful in isolating problematic tests that cause the test host to crash. When a crash is detected, it creates a sequence file in TestResults/<Guid>/<Guid>_Sequence.xml that captures the order of tests that were run before the crash.

This option does not create a memory dump and is not helpful when the test is hanging.
--blame-crash (Available since .NET 5.0 SDK)

Runs the tests in blame mode and collects a crash dump when the test host exits unexpectedly. This option depends on the version of .NET used, the type of error, and the operating system.

For exceptions in managed code, a dump will be automatically collected on .NET 5.0 and later versions. It will generate a dump for testhost or any child process that also ran on .NET 5.0 and crashed. Crashes in native code will not generate a dump. This option works on Windows, macOS, and Linux.

Crash dumps in native code, or when using .NET Core 3.1 or earlier versions, can only be collected on Windows, by using Procdump. A directory that contains procdump.exe and procdump64.exe must be in the PATH or PROCDUMP_PATH environment variable. Download the tools. Implies --blame.

To collect a crash dump from a native application running on .NET 5.0 or later, the usage of Procdump can be forced by setting the VSTEST_DUMP_FORCEPROCDUMP environment variable to 1.
--blame-crash-dump-type <DUMP_TYPE> (Available since .NET 5.0 SDK)

The type of crash dump to be collected. Supported dump types are full (default), and mini. Implies --blame-crash.
--blame-crash-collect-always (Available since .NET 5.0 SDK)

Collects a crash dump on expected as well as unexpected test host exit.
--blame-hang (Available since .NET 5.0 SDK)

Run the tests in blame mode and collects a hang dump when a test exceeds the given timeout.
--blame-hang-dump-type <DUMP_TYPE> (Available since .NET 5.0 SDK)

The type of crash dump to be collected. It should be full, mini, or none. When none is specified, test host is terminated on timeout, but no dump is collected. Implies --blame-hang.
--blame-hang-timeout <TIMESPAN> (Available since .NET 5.0 SDK)

Per-test timeout, after which a hang dump is triggered and the test host process and all of its child processes are dumped and terminated. The timeout value is specified in one of the following formats:
- 1.5h, 1.5hour, 1.5hours
- 90m, 90min, 90minute, 90minutes
- 5400s, 5400sec, 5400second, 5400seconds
- 5400000ms, 5400000mil, 5400000millisecond, 5400000milliseconds
When no unit is used (for example, 5400000), the value is assumed to be in milliseconds. When used together with data driven tests, the timeout behavior depends on the test adapter used. For xUnit, NUnit. and MSTest 2.2.4+, the timeout is renewed after every test case. For MSTest before version 2.2.4, the timeout is used for all test cases. This option is supported on Windows with netcoreapp2.1 and later, on Linux with netcoreapp3.1 and later, and on macOS with net5.0 or later. Implies --blame and --blame-hang.

-c|--configuration <CONFIGURATION>

Defines the build configuration. The default for most projects is Debug, but you can override the build configuration settings in your project.

--collect <DATA_COLLECTOR_NAME>

Enables data collector for the test run. For more information, see Monitor and analyze test run.

For example you can collect code coverage by using the --collect "Code Coverage" option. For more information, see Use code coverage, Customize code coverage analysis, and GitHub issue dotnet/docs#34479.

To collect code coverage you can also use Coverlet by using the --collect "XPlat Code Coverage" option.
-d|--diag <LOG_FILE>

Enables diagnostic mode for the test platform and writes diagnostic messages to the specified file and to files next to it. The process that is logging the messages determines which files are created, such as *.host_<date>.txt for test host log, and *.datacollector_<date>.txt for data collector log.
-e|--environment <NAME="VALUE">

Sets the value of an environment variable. Creates the variable if it does not exist, overrides if it does exist. Use of this option will force the tests to be run in an isolated process. The option can be specified multiple times to provide multiple variables.
-f|--framework <FRAMEWORK>

The target framework moniker (TFM) of the target framework to run tests for. The target framework must also be specified in the project file.
--filter <EXPRESSION>

Filters tests in the current project using the given expression. Only tests that match the filter expression are run. For more information, see the Filter option details section. For more information and examples on how to use selective unit test filtering, see Running selective unit tests.

-?|-h|--help

Prints out a description of how to use the command.

--interactive

Allows the command to stop and wait for user input or action. For example, to complete authentication. Available since .NET Core 3.0 SDK.

-l|--logger <LOGGER>

Specifies a logger for test results and optionally switches for the logger. Specify this parameter multiple times to enable multiple loggers. For more information, see Reporting test results, Switches for loggers, and the examples later in this article.

In order to pass command-line switches to the logger:
- Use the full name of the switch, not the abbreviated form (for example, verbosity instead of v).
- Omit any leading dashes.
- Replace the space separating each switch with a semicolon ;.
- If the switch has a value, replace the colon separator between that switch and its value with the equals sign =.
For example, -v:detailed --consoleLoggerParameters:ErrorsOnly would become verbosity=detailed;consoleLoggerParameters=ErrorsOnly.
--no-build

Doesn't build the test project before running it. It also implicitly sets the --no-restore flag.
--nologo

Run tests without displaying the Microsoft TestPlatform banner. Available since .NET Core 3.0 SDK.
--no-restore

Doesn't execute an implicit restore when running the command.
-o|--output <OUTPUT_DIRECTORY>

Directory in which to find the binaries to run. If not specified, the default path is ./bin/<configuration>/<framework>/. For projects with multiple target frameworks (via the TargetFrameworks property), you also need to define --framework when you specify this option. dotnet test always runs tests from the output directory. You can use AppDomain.BaseDirectory to consume test assets in the output directory.
- .NET 7.0.200 SDK and later
  
  If you specify the --output option when running this command on a solution, the CLI will emit a warning (an error in 7.0.200) due to the unclear semantics of the output path. The --output option is disallowed because all outputs of all built projects would be copied into the specified directory, which isn't compatible with multi-targeted projects, as well as projects that have different versions of direct and transitive dependencies. For more information, see Solution-level --output option no longer valid for build-related commands.

--os <OS>

Specifies the target operating system (OS). This is a shorthand syntax for setting the Runtime Identifier (RID), where the provided value is combined with the default RID. For example, on a win-x64 machine, specifying --os linux sets the RID to linux-x64. If you use this option, don't use the -r|--runtime option. Available since .NET 6.

--results-directory <RESULTS_DIR>

The directory where the test results are going to be placed. If the specified directory doesn't exist, it's created. The default is TestResults in the directory that contains the project file.

Short form -r available in .NET SDK versions earlier than 7.
-r|--runtime <RUNTIME_IDENTIFIER>

The target runtime to test for.

Short form -r available starting in .NET SDK 7.
-s|--settings <SETTINGS_FILE>

The .runsettings file to use for running the tests. The TargetPlatform element (x86|x64) has no effect for dotnet test. To run tests that target x86, install the x86 version of .NET Core. The bitness of the dotnet.exe that is on the path is what will be used for running tests. For more information, see the following resources:
- Configure unit tests by using a .runsettings file.
- Configure a test run
-t|--list-tests

List the discovered tests instead of running the tests.

-v|--verbosity <LEVEL>

Sets the verbosity level of the command. Allowed values are q[uiet], m[inimal], n[ormal], d[etailed], and diag[nostic]. The default is minimal. For more information, see LoggerVerbosity.

args

Specifies extra arguments to pass to the adapter. Use a space to separate multiple arguments.

The list of possible arguments depends upon the specified behavior:
- When you specify a project, solution, or a directory, or if you omit this argument, the call is forwarded to msbuild. In that case, the available arguments can be found in the dotnet msbuild documentation.
- When you specify a .dll or an .exe, the call is forwarded to vstest. In that case, the available arguments can be found in the dotnet vstest documentation.
RunSettings arguments

Inline RunSettings are passed as the last arguments on the command line after "-- " (note the space after --). Inline RunSettings are specified as [name]=[value] pairs. A space is used to separate multiple [name]=[value] pairs.

Example: dotnet test -- MSTest.DeploymentEnabled=false MSTest.MapInconclusiveToFailed=True

For more information, see Passing RunSettings arguments through command line.

Examples

Run the tests in the project in the current directory:
```
dotnet test
```

Run the tests in the test1 project:

dotnet test ~/projects/test1/test1.csproj

Run the tests using test1.dll assembly:

dotnet test ~/projects/test1/bin/debug/test1.dll

Run the tests in the project in the current directory, and generate a test results file in the trx format:
```
dotnet test --logger trx
```
Run the tests in the project in the current directory, and generate a code coverage file (after installing Coverlet collectors integration):
```
dotnet test --collect:"XPlat Code Coverage"
```
Run the tests in the project in the current directory, and generate a code coverage file (Windows only):
```
dotnet test --collect "Code Coverage"
```
Run the tests in the project in the current directory, and log with detailed verbosity to the console:
```
dotnet test --logger "console;verbosity=detailed"
```
Run the tests in the project in the current directory, and log with the trx logger to testResults.trx in the TestResults folder:
```
dotnet test --logger "trx;logfilename=testResults.trx"
```
Since the log file name is specified, the same name is used for each target framework in the case of a multi-targeted project. The output for each target framework overwrites the output for preceding target frameworks. The file is created in the TestResults folder in the test project folder, because relative paths are relative to that folder. The following example shows how to produce a separate file for each target framework.
Run the tests in the project in the current directory, and log with the trx logger to files in the TestResults folder, with file names that are unique for each target framework:
```
dotnet test --logger:"trx;LogFilePrefix=testResults"
```
Run the tests in the project in the current directory, and log with the html logger to testResults.html in the TestResults folder:
```
dotnet test --logger "html;logfilename=testResults.html"
```
Run the tests in the project in the current directory, and report tests that were in progress when the test host crashed:
```
dotnet test --blame
```
Run the tests in the test1 project, providing the -bl (binary log) argument to msbuild:
```
dotnet test ~/projects/test1/test1.csproj -bl
```
Run the tests in the test1 project, setting the MSBuild DefineConstants property to DEV:
```
dotnet test ~/projects/test1/test1.csproj -p:DefineConstants="DEV"
```
Run the tests in the test1 project, setting the MSBuild TestTfmsInParallel property to false:
```
dotnet test ~/projects/test1/test1.csproj -p:TestTfmsInParallel=false
```

--filter <EXPRESSION>

<Expression> has the format <property><operator><value>[|&<Expression>].

<property> is an attribute of the Test Case. The following are the properties supported by popular unit test frameworks:

Test Framework	Supported properties
MSTest	FullyQualifiedName Name ClassName Priority TestCategory
xUnit	FullyQualifiedName DisplayName Category
NUnit	FullyQualifiedName Name Category Priority

The <operator> describes the relationship between the property and the value:

Operator	Function
`=`	Exact match
`!=`	Not exact match
`~`	Contains
`!~`	Not contains

<value> is a string. All the lookups are case insensitive.

An expression without an <operator> is automatically considered as a contains on FullyQualifiedName property (for example, dotnet test --filter xyz is same as dotnet test --filter FullyQualifiedName~xyz).

Expressions can be joined with conditional operators:

Operator	Function
`\|`	OR
`&`	AND

You can enclose expressions in parenthesis when using conditional operators (for example, (Name~TestMethod1) | (Name~TestMethod2)).

For more information and examples on how to use selective unit test filtering, see Running selective unit tests.

Share via