FAKE - F# Make - A DSL for build tasks and more FAKE
6.0.0
Edit this page

Lookig for older versions of the documentation, pre FAKE v6? See 

v5.fake.build

Getting Started

Table of Content:

In this tutorial you will learn how to set up a complete build infrastructure with FAKE. This includes: how to install and use the latest FAKE version, how to edit and run scripts, how to automatically compile your C# or F# projects, and finally how to automatically run NUnit tests on your projects.

If you are interested in what FAKE actually is, see our FAQ.

But before going into these details, let's first discuss the possible ways you can use to run FAKE.

Different ways to run FAKE

FAKE is built with modularity in mind, in which it doesn't constrain you with one way to write or run your build scripts. Over the years, FAKE evolves and allowed users to consume and run FAKE in multiple ways, the currently used ways are:

Let's discuss each approach.

Run FAKE using FAKE runner

The simplest way to run a build script is to use the FAKE runner. The FAKE runner was the de-facto for running build scripts in FAKE, because it offers multiple features that were not possible before, including automatic resolution of dependencies using NuGet or Paket dependency managers and caching which makes FAKE super fast, among others.

Assuming you have already installed FAKE runner (please see Install FAKE Runner section below on how to install FAKE runner). To get started with FAKE runner, create an F# interactive script (build.fsx):

#r "paket:
nuget Fake.Core.Target //"
// include Fake modules, see Fake modules section

open Fake.Core

// *** Define Targets ***
Target.create "Hello" (fun _ ->
  printfn "hello from FAKE!"
)

// *** Start Build ***
Target.runOrDefault "Hello"

Next, in a command line interface, navigate to the script directory and enter the following command (assuming you installed FAKE runner as a global .Net tool):

fake build

You should see the message hello from FAKE! printed on the screen! You have ran your first script in FAKE!

The above script uses Paket dependency manager to resolve and download FAKE's tagret module, which allows you to define targets and organize your script. You can check targets module documentation for more info. After resolving the target module, the script open the module and define a tagret named hello which prints a message to standard output. Finally, the final statement in the script is the entry point of your script which designate the Hello target as the default target to run.

The automatic resolution of dependencies, caching and many other features are provided by the FAKE runner under the hood to make running your script a breeze.

Run FAKE using F# interactive (FSI)

Next, let's discuss the second way, which it utilizing FSI to run our script. FAKE runner was written to provide features that were not available back then in FSI, including the dependencies resolution using a dependency manager such as NuGet or Paket. However, FSI has evolved as well and now it offers these capabilities.

From the previous way, note that the script is just a plain F# interactive file! Nothing special about it! However, to successfully run this script using FSI we need to tell FSI that it need to use Paket dependency manager to resolve the dependency we specified (FAKE's tagret module). Paket has step-by-step guide on how to let FSI knows about it, please see Paket's F# Interactive Integration Guide.

After configuring Paket, now you are ready to run your script, enter the following command to run it:

dotnet si build.fsx --compilertool:"~/.nuget/packages/fsharp.dependencymanager.paket/6.0.0-alpha055/lib/netstandard2.0"

Run FAKE using a dedicated build project

The last method we will discuss is using a dedicated build project to host your build script and its dependencies. The steps you need to take include:

This method utilize .Net project managment and dependecy resolution and offers you possibility to stay within project based approach. The build project is a normal .Net console application that you can manage dependencies to it using .Net command line, and run it also using .Net command line.

The advantages of this approach is that, it saves you from cases in which FAKE runner or FSI may encouner running your build script. Specifically, dependency resolution and when upgrading to a newer version of .Net SDk. This approach is becoming incresingly adopted in FAKE community. Some of the projects that are using this approach are:

You can examine these two repository to get to know the approach and how to implement it in your projects.

Thanks to @baronfel and @aboy021 for proposing this solution and discussing it in this StackOverflow question: Fix for "Package manager key paket was not registered" in build.fsx

Install FAKE Runner

Now that you know the possible ways to run FAKE scripts, let's zoom in and see how to install the FAKE runner. FAKE is completely written in F# and all build scripts will also be written in F#, but this doesn't imply that you have to learn programming in F#. In fact the FAKE syntax is hopefully very easy to learn.

There are various ways to install FAKE:

  1. Install FAKE as a local dotnet tool (easiest, but needs .NET Core SDK Version 3 or newer): First you need to create a tool manifest, which should be committed to your repository:

    dotnet new tool-manifest

    To install FAKE, run:

    dotnet tool install fake-cli

    Use --version to display the version of FAKE. See the local_tool branch of fake-bootstrap for ideas to bootstrap in your CI process. To run fake use dotnet fake build. To restore/download fake on another machine use dotnet tool restore after this command dotnet fake build should work as expected.

  2. Install FAKE as a global dotnet tool (easiest, but needs .NET Core SDK): To install FAKE globally, run:

    dotnet tool install fake-cli -g

    To install FAKE into your_tool_path, run:

    dotnet tool install fake-cli --tool-path your_tool_path

    Use --version to display the version of FAKE. See the global_tool branch of fake-bootstrap for ideas to bootstrap in your CI process. If you run into issues, see this.

  3. Bootstrap via the fake dotnet new template. The template bootstraps FAKE and sets up a basic build-script. To install the template run:

    dotnet new -i "fake-template::*"

    Then run the template with:

    dotnet new fake

    See the template page for more information. Install the fake or fake-netcore package for your system ( currently chocolatey). Example choco install fake. We also provide a Debian package at the releases-page

  4. Download the portable zip. We distribute a .NET Core version of FAKE without the .NET Core runtime. This version assumes an existing dotnet sdk installation while the non-portable installation doesn't.

    Just use the -portable version of the downloads, extract it and execute.

    dotnet fake.dll

    The advantage of this method is that it is portable (ie. distribute the same binaries) and requires less bandwidth. The disadvantage is that you need to have a dotnet sdk installed and that you need to prefix all calls with dotnet /full/path/to/fake.dll <args> which is equal to fake <args> in other installation methods.

  5. Download the runtime specific zip. Just use the -<runtime> version matching your specific platform of the downloads (for example fake-dotnetcore-win7-x64.zip), extract it and execute the fake binary. Add the extracted binary to your PATH to just execute fake from any directory.

  6. Bootstrap via a shell script (fake.cmd/fake.sh), see this example project

    These scripts have no versioning story. You either need to take care of versions yourself (and lock them) or your builds might break on major releases.

    now you can use

    fake --help

    This is basically it. You can now execute fake commands.

One note on Intellisense

This section is to clarify when and how the intellisense updates when you add new modules (short form: Delete the <script>.fsx.lock file and re-run fake to update all files).

We recommend Visual Studio Code with the Ionide extension for best FAKE tooling support, including proper debugging.

The warning FS0988: Main module of program is empty: nothing will happen when it is run indicates that you have not written any code into the script yet.

If you change your dependencies you need to delete myscript.fsx.lock and run the script again for intellisense to update. Intellisense is shown for the full framework while the script is run as netcoreapp20 therefore intellisense might show APIs which are not actually usable.

Examples

This section offer detailed steps to write a FAKE script to compile, test and run a .Net application. We will start with a meinimal example as a refresh for FAKE syntax and script structure. Then discuss the .Net application build script, and finally, we will close this section with some advanced examples of FAKE scripts. Let's dive in.

Minimal Example

Create a file named build.fsx with the following contents:

#r "paket:
nuget Fake.Core.Target //"
// include Fake modules, see Fake modules section

open Fake.Core

// *** Define Targets ***
Target.create "Clean" (fun _ ->
  Trace.log " --- Cleaning stuff --- "
)

Target.create "Build" (fun _ ->
  Trace.log " --- Building the app --- "
)

Target.create "Deploy" (fun _ ->
  Trace.log " --- Deploying app --- "
)

open Fake.Core.TargetOperators

// *** Define Dependencies ***
"Clean"
  ==> "Build"
  ==> "Deploy"

// *** Start Build ***
Target.runOrDefault "Deploy"

This is the minimal FAKE script, which has the following major parts: 1. The #r directive at the top define the dependencies the project need. 2. The // *** Define Targets *** section define targets using FAKE target module. These targets will hold your build logic. 3. The // *** Define Dependencies *** section define the dependencies between your targets. In this example we have specified that to run Deploy, Build must be run first, and to run Build, Clean must be run first. So order is Clean then Build then Deploy. 4. The final statements designate the Deploy target as the default target to run.

Run this file by executing:

fake run build.fsx

Compiling and Building your .NET Application

This example will guide you by adding a fake script to your existing .NET application.

Getting Started

Initially we need to create a file called build.fsx where all our build-logic will reside. Create a new file with Visual Studio or Visual Studio Code (with ionide) and paste the following content:

#r "paket:
nuget Fake.Core.Target //"
#load "./.fake/build.fsx/intellisense.fsx"

This is all we need for now to declare that we need the Fake.Core.Target module and want to enable intellisense.

Now run fake run build.fsx to make fake prepare our environment. Now our IDE can load the dependencies and we will have intellisense enabled (you might need to reopen the script file on some editors).

Now that we have setup our basic environment to edit the script file we can add our first target:

#r "paket:
nuget Fake.Core.Target //"
#load "./.fake/build.fsx/intellisense.fsx"

open Fake.Core

// Default target
Target.create "Default" (fun _ ->
  Trace.trace "Hello World from FAKE"
)

// start build
Target.runOrDefault "Default"

As you can see the code is really simple. The first few lines (nuget Fake.Core.Target and open Fake.Core) load the fake modules we need and this is vital for all build scripts to support creating and running targets. The #load line is optional but a good way to make the IDE aware of all the modules (for intellisense and IDE support)

After this header the Default target is defined. A target definition contains two important parts. The first is the name of the target (here "Default") and the second is an action (here a simple trace of "Hello world").

The last line runs the "Default" target - which means it executes the defined action of the target.

Try running your new target via fake run build.fsx or the shortcut for a file called build.fsx: fake build

Cleaning the Last Build Output

A typical first step in most build scenarios is to clean the output of the last build. We can achieve this in two steps:

First change your header to the following by adding the Fake.IO.FileSystem module:

#r "paket:
nuget Fake.IO.FileSystem
nuget Fake.Core.Target //"
#load "./.fake/build.fsx/intellisense.fsx"

Now we need to remove the build.fsx.lock file and run fake build in order to restore the newly added Fake.IO.FileSystem module.

Since we now can rely on intellisense we can easily discover the various modules and functions in Fake.IO, for example the Shell module provides various functions you expect from regular shell scripting, we will use Shell.cleanDir which will ensure the given directory is empty by deleting everything within or creating the directory if required:

#r "paket:
nuget Fake.IO.FileSystem
nuget Fake.Core.Target //"
#load "./.fake/build.fsx/intellisense.fsx"

open Fake.Core
open Fake.IO

// Properties
let buildDir = "./build/"

// Targets
Target.create "Clean" (fun _ ->
  Shell.cleanDir buildDir
)

Target.create "Default" (fun _ ->
  Trace.trace "Hello World from FAKE"
)

// Dependencies
open Fake.Core.TargetOperators

"Clean"
  ==> "Default"

// start build
Target.runOrDefault "Default"

You can explore the APIs for example by writing Fake.IO. and waiting for intellisense (or pressing Ctrl+Space). You can remove Fake.IO once you put open Fake.IO on top.

We introduced some new concepts in this snippet. At first we defined a global property called buildDir with the relative path of a temporary build folder.

In the Clean target we use the Shell.cleanDir task to clean up this build directory. As explained above this simply deletes all files in the folder or creates the directory if necessary.

In the dependencies section we say that the Default target has a dependency on the Clean target. In other words Clean is a prerequisite of Default and will run before the Default target is executed:

alt text

Compiling the Application

In the next step we want to compile our C# libraries, which means we want to compile all csproj-files under /src/app with MSBuild.

Again we need a new module for this, namely Fake.DotNet.MSBuild.

Just like before add the required module on top via nuget Fake.DotNet.MSBuild, delete the build.fsx.lock file and run the script. Now edit the script so it looks like this:

#r "paket:
nuget Fake.IO.FileSystem
nuget Fake.DotNet.MSBuild
nuget Fake.Core.Target //"
#load "./.fake/build.fsx/intellisense.fsx"

open Fake.IO
open Fake.IO.Globbing.Operators //enables !! and globbing
open Fake.DotNet
open Fake.Core

// Properties
let buildDir = "./build/"

// Targets
Target.create "Clean" (fun _ ->
  Shell.cleanDir buildDir
)

Target.create "BuildApp" (fun _ ->
  !! "src/app/**/*.csproj"
    |> MSBuild.runRelease id buildDir "Build"
    |> Trace.logItems "AppBuild-Output: "
)

Target.create "Default" (fun _ ->
  Trace.trace "Hello World from FAKE"
)

open Fake.Core.TargetOperators

"Clean"
  ==> "BuildApp"
  ==> "Default"

// start build
Target.runOrDefault "Default"

We defined a new build target named "BuildApp" which compiles all csproj-files with the MSBuild task and the build output will be copied to buildDir.

In order to find the right project files FAKE scans the folder src/app/ and all subfolders with the given pattern ( the !! operator was imported from Fake.IO.FileSystem via open Fake.IO.Globbing.Operators). Therefore a similar FileSet definition like in NAnt or MSBuild (see project page for details) is used.

In addition the target dependencies are extended again. Now Default is dependent on BuildApp and BuildApp needs Clean as a prerequisite.

This means the execution order is: Clean ==> BuildApp ==> Default.

alt text

Compiling Test Projects

Now our main application will be built automatically and it's time to build the test project. We use the same concepts as before:

#r "paket:
nuget Fake.IO.FileSystem
nuget Fake.DotNet.MSBuild
nuget Fake.Core.Target //"
#load "./.fake/build.fsx/intellisense.fsx"

open Fake.IO
open Fake.IO.Globbing.Operators
open Fake.DotNet
open Fake.Core


// Properties
let buildDir = "./build/"
let testDir  = "./test/"

// Targets
Target.create "Clean" (fun _ ->
  Shell.cleanDirs [buildDir; testDir]
)

Target.create "BuildApp" (fun _ ->
    !! "src/app/**/*.csproj"
    |> MSBuild.runRelease id buildDir "Build"
    |> Trace.logItems "AppBuild-Output: "
)

Target.create "BuildTest" (fun _ ->
  !! "src/test/**/*.csproj"
    |> MSBuild.runDebug id testDir "Build"
    |> Trace.logItems "TestBuild-Output: "
)

Target.create "Default" (fun _ ->
  Trace.trace "Hello World from FAKE"
)

open Fake.Core.TargetOperators
"Clean"
  ==> "BuildApp"
  ==> "BuildTest"
  ==> "Default"

// start build
Target.runOrDefault "Default"

This time we defined a new target "BuildTest" which compiles all C# projects below src/test/ in Debug mode and we put the target into our build order.

Running the Tests with NUnit

Now all our projects will be compiled and we can use FAKE's NUnit task in order to let NUnit test our assembly (we have to add a new module for this: Fake.DotNet.Testing.NUnit):

#r "paket:
nuget Fake.IO.FileSystem
nuget Fake.DotNet.MSBuild
nuget Fake.DotNet.Testing.NUnit
nuget Fake.Core.Target //"
#load "./.fake/build.fsx/intellisense.fsx"

open Fake.IO
open Fake.IO.Globbing.Operators
open Fake.DotNet
open Fake.DotNet.Testing
open Fake.Core

// Properties
let buildDir = "./build/"
let testDir  = "./test/"

// Targets
Target.create "Clean" (fun _ ->
    Shell.CleanDirs [buildDir; testDir]
)

Target.create "BuildApp" (fun _ ->
   !! "src/app/**/*.csproj"
     |> MSBuild.runRelease id buildDir "Build"
     |> Trace.logItems "AppBuild-Output: "
)

Target.create "BuildTest" (fun _ ->
    !! "src/test/**/*.csproj"
      |> MSBuild.runDebug id testDir "Build"
      |> Trace.logItems "TestBuild-Output: "
)

Target.create "Test" (fun _ ->
    !! (testDir + "/NUnit.Test.*.dll")
      |> NUnit3.run (fun p ->
          {p with
                ShadowCopy = false })
)

Target.create "Default" (fun _ ->
    Trace.trace "Hello World from FAKE"
)

// Dependencies
open Fake.Core.TargetOperators
"Clean"
  ==> "BuildApp"
  ==> "BuildTest"
  ==> "Test"
  ==> "Default"

// start build
Target.runOrDefault "Default"

Our new Test target scans the test directory for test assemblies and runs them with the NUnit runner. FAKE automatically tries to locate the runner in one of your subfolders. See the NUnit3 task documentation if you need to specify the tool path explicitly.

The mysterious part (fun p -> ...) simply overrides the default parameters of the NUnit task and allows to specify concrete parameters.

alt text

Advanced Examples

You can check these repositories for advanced usage of FAKE: - See FAKE build.fsx - See SAFE-Stack build.fsx - See csharptofsharp build.fsx

Runtime Assemblies in FAKE Runner

FAKE runner uses .Net6 runtime assemblies when compiling and running a script.

Get to Know FAKE CLI

See Fake command line for more information on command line and available commands and options

What's Next?