Mill


Table of Contents

Contrib Modules

Mill InternalsThirdparty Modules

The plugins in this section are developed/maintained in the mill git tree.

When using one of these, it is important that the versions you load match your mill version. To facilitate this, Mill will automatically replace the $MILL_VERSION literal in your ivy imports with the correct value.

For instance :

import $ivy`com.lihaoyi::mill-contrib-bloop:$MILL_VERSION`

Bloop

This plugin generates bloop configuration from your build file, which lets you use the bloop CLI for compiling, and makes your scala code editable in Metals

Quickstart

// build.sc (or any other .sc file it depends on, including predef)
import $ivy.`com.lihaoyi::mill-contrib-bloop:$MILL_VERSION`

Then in your terminal :

> mill mill.contrib.Bloop/install

It generate correct bloop config for any JavaModule, ScalaModule, ScalaJsModule or ScalaNativeModule under the .bloop folder

Mix-in

You can mix-in the Bloop.Module trait with any JavaModule to quickly access the deserialised configuration for that particular module:

// build.sc
import mill._
import mill.scalalib._
import mill.contrib.Bloop

object MyModule extends ScalaModule with Bloop.Module {
  def myTask = T { bloop.config() }
}

Note regarding metals

Generating the bloop config should be enough for metals to pick it up and for features to start working in vscode (or the bunch of other editors metals supports). However, note that this applies only to your project sources. Your mill/ammonite related .sc files are not yet supported by metals.

The generated bloop config references the semanticDB compiler plugin required by metals to function. If need be, the version of semanticDB can be overriden by extending mill.contrib.bloop.BloopImpl in your own space.

Note regarding current mill support in bloop

The mill-bloop integration currently present in the bloop codebase will be deprecated in favour of this implementation.

BuildInfo

Generate scala code from your buildfile. This plugin generates a single object containing information from your build.

To declare a module that uses BuildInfo you must extend the mill.contrib.buildinfo.BuildInfo trait when defining your module.

Quickstart:

// build.sc
import $ivy.`com.lihaoyi::mill-contrib-buildinfo:$MILL_VERSION`
import mill.contrib.buildinfo.BuildInfo

object project extends BuildInfo {
  val name = "poject-name"
  def  buildInfoMembers: T[Map[String, String]] = T {
    Map(
      "name" -> name),
      "scalaVersion" -> scalaVersion()
    )
  }
}

Configuration options

Docker

Automatically build docker images from your mill project.

Requires the docker CLI to be installed.

In the simplest configuration just extend DockerModule and declare a DockerConfig object.

import mill._, scalalib._

import ivy`com.lihaoyi::mill-contrib-docker:VERSION`
import contrib.docker.DockerModule

object foo extends JavaModule with DockerModule {
  object docker extends DockerConfig
}

Then

$ mill foo.docker.build
$ docker run foo

Configuration

Configure the image by overriding tasks in the DockerConfig object

object docker extends DockerConfig {
  // Override tags to set the output image name
  def tags = List("aws_account_id.dkr.ecr.region.amazonaws.com/hello-repository")

  def baseImage = "openjdk:11"

  // Configure whether the docker build should check the remote registry for a new version of the base image before building.
  // By default this is true if the base image is using a latest tag
  def pullBaseImage = true
}

Run mill in interactive mode to see the docker client output, like mill -i foo.docker.build.

Flyway

Enables you to configure and run Flyway commands from your mill build file. The flyway module currently supports the most common flyway use cases with file based migrations.

Configure flyway by overriding settings in your module. For example

// build.sc

import mill._, scalalib._

import ivy`com.lihaoyi::mill-contrib-flyway:$MILL_VERSION`
import contrib.flyway.FlywayModule

object foo extends ScalaModule with FlywayModule {
  def scalaVersion = "2.12.8"

  //region flyway
  def flywayUrl = "jdbc:postgresql:myDb" // required
  def flywayDriverDeps = Agg(ivy"org.postgresql:postgresql:42.2.5") // required
  def flywayUser = "postgres" // optional
  // def flywayPassword = "" // optional
  //endregion
}

Flyway will look for migration files in db/migration in all resources folders by default. This should work regardless of if you are using a mill or sbt project layout.

You can then run common flyway commands like

mill foo.flywayClean
mill foo.flywayInfo
mill foo.flywayMigrate

REMINDER: You should never hard-code credentials or check them into a version control system. You should write some code to populate the settings for flyway instead. For example def flywayPassword = T.input(T.ctx().env("FLYWAY_PASSWORD"))

Play Framework

This module adds basic Play Framework support to mill:

There is no specific Play Java support, building a Play Java application will require a bit of customization (mostly adding the proper dependencies).

Using the plugin

There are 2 base modules and 2 helper traits in this plugin, all of which can be found in mill.playlib.

The base modules:

The two helper traits:

Using PlayModule

In order to use the PlayModule for your application, you need to provide the scala, Play and Twirl versions. You also need to define your own test object which extends the provided PlayTests trait.

// build.sc
import mill._
import $ivy.`com.lihaoyi::mill-contrib-playlib:$MILL_VERSION`,  mill.playlib._


object core extends PlayModule {
    //config
    override def scalaVersion= T{"2.12.8"}
    override def playVersion= T{"2.7.0"}
    override def twirlVersion= T{"1.4.0"}

    object test extends PlayTests
}

Using the above definition, your build will be configured to use the default Play layout:

.
├── build.sc
└── core
    ├── app
    │   ├── controllers
    │   └── views
    ├── conf
    │   └── application.conf
    │   └── routes
    │   └── ...
    ├── logs
    ├── public
    │   ├── images
    │   ├── javascripts
    │   └── stylesheets
    └── test
        └── controllers

The following compile dependencies will automatically be added to your build:

ivy"com.typesafe.play::play:${playVersion()}",
ivy"com.typesafe.play::play-guice:${playVersion()}",
ivy"com.typesafe.play::play-server:${playVersion()}",
ivy"com.typesafe.play::play-logback:${playVersion()}"

Scala test will be setup as the default test framework and the following test dependencies will be added (the actual version depends on the version of Play you are pulling 2.6.x or 2.7.x):

ivy"org.scalatestplus.play::scalatestplus-play::4.0.1"

In order to have a working start command the following runtime dependency is also added:

ivy"com.typesafe.play::play-akka-http-server:${playVersion()}"

Using PlayApiModule

The PlayApiModule trait behaves the same as the PlayModule trait but it won't process .scala .html files and you don't need to define the `twirlVersion:

// build.sc
import mill._
import $ivy.`com.lihaoyi::mill-contrib-playlib:$MILL_VERSION`,  mill.playlib._


object core extends PlayApiModule {
    //config
    override def scalaVersion= T{"2.12.8"}
    override def playVersion= T{"2.7.0"}

    object test extends PlayTests
}

Play configuration options

The Play modules themselves don't have specific configuration options at this point but the router module configuration options and the Twirl module configuration options are applicable.

Additional play libraries

The following helpers are available to provide additional Play Framework dependencies:

If you want to add an optional library using the helper you can do so by overriding ivyDeps like in the following example build:

// build.sc
import mill._
import $ivy.`com.lihaoyi::mill-contrib-playlib:$MILL_VERSION`,  mill.playlib._


object core extends PlayApiModule {
    //config
    override def scalaVersion= T{"2.12.8"}
    override def playVersion= T{"2.7.0"}

    object test extends PlayTests

    override def ivyDeps = T{ super.ivyDeps() ++ Agg(ws(), filters()) }
}

Commands equivalence

Mill commands are targets on a named build. For example if your build is called core:

Using SingleModule

The SingleModule trait allows you to have the build descriptor at the same level as the source code on the filesystem. You can move from there to a multi-module build either by refactoring your directory layout into multiple subdirectories or by using mill's nested modules feature.

Looking back at the sample build definition in Using PlayModule:

// build.sc
import mill._
import $ivy.`com.lihaoyi::mill-contrib-playlib:$MILL_VERSION`,  mill.playlib._


object core extends PlayModule {
    //config
    override def scalaVersion= T{"2.12.8"}
    override def playVersion= T{"2.7.0"}
    override def twirlVersion= T{"1.4.0"}

    object test extends PlayTests
}

The directory layout was:

.
├── build.sc
└── core
    ├── app
    │   ├── controllers
    │   └── views
    ├── conf
    │   └── application.conf
    │   └── routes
    │   └── ...
    ├── logs
    ├── public
    │   ├── images
    │   ├── javascripts
    │   └── stylesheets
    └── test
        └── controllers

by mixing in the SingleModule trait in your build:

// build.sc
import mill._
import $ivy.`com.lihaoyi::mill-contrib-playlib:$MILL_VERSION`,  mill.playlib._


object core extends PlayModule with SingleModule {
	//config
	override def scalaVersion= T{"2.12.8"}
	override def playVersion= T{"2.7.0"}
	override def twirlVersion= T{"1.4.0"}

	object test extends PlayTests
}

the layout becomes:

.
└── core
    ├── build.sc
    ├── app
    │   ├── controllers
    │   └── views
    ├── conf
    │   └── application.conf
    │   └── routes
    │   └── ...
    ├── logs
    ├── public
    │   ├── images
    │   ├── javascripts
    │   └── stylesheets
    └── test
        └── controllers

Using the router module directly

If you want to use the router module in a project which doesn't use the default Play layout, you can mix-in the mill.playlib.routesModule trait directly when defining your module. Your app must define playVersion and scalaVersion.

// build.sc
import mill._
import $ivy.`com.lihaoyi::mill-contrib-playlib:$MILL_VERSION`,  mill.playlib._


object app extends ScalaModule with RouterModule {
  def playVersion= T{"2.7.0"}
  def scalaVersion= T{"2.12.8"}
}
Router Configuration options
Details

The following filesystem layout is expected by default:

.
├── app
│   └── routes
│       └── routes
└── build.sc

RouterModule adds the compileRouter task to the module:

mill app.compileRouter

(it will be automatically run whenever you compile your module)

This task will compile routes templates into the out/app/compileRouter/dest directory. This directory must be added to the generated sources of the module to be compiled and made accessible from the rest of the code. This is done by default in the trait, but if you need to have a custom override for generatedSources you can get the list of files from routerClasses

To add additional imports to all of the routes:

// build.sc
import mill.scalalib._

import $ivy.`com.lihaoyi::mill-contrib-playlib:$MILL_VERSION`,  mill.playlib._

object app extends ScalaModule with RouterModule {
  def playVersion = "2.7.0"
  override def routesAdditionalImport = Seq("my.additional.stuff._", "my.other.stuff._")
}

ScalaPB

This module allows ScalaPB to be used in Mill builds. ScalaPB is a Protocol Buffers compiler plugin that generates Scala case classes, encoders and decoders for protobuf messages.

To declare a module that uses ScalaPB you can extend the mill.contrib.scalapblib.ScalaPBModule trait when defining your module.

This creates a Scala module which compiles .proto files in the protobuf folder of the module with ScalaPB and adds the resulting .scala sources to your module's generatedSources.

// build.sc

import $ivy.`com.lihaoyi::mill-contrib-scalapblib:$MILL_VERSION`
import contrib.scalapblib._

object example extends ScalaPBModule {
  def scalaVersion = "2.12.6"
  def scalaPBVersion = "0.7.4"
}

This defines a project with the following layout:

build.sc
example/
    src/
    protobuf/
    resources/

Configuration options

If you'd like to configure the options that are passed to the ScalaPB compiler directly, you can override the scalaPBOptions task, for example:

// build.sc

import $ivy.`com.lihaoyi::mill-contrib-scalapblib:$MILL_VERSION`
import contrib.scalapblib._

object example extends ScalaPBModule {
  def scalaVersion = "2.12.6"
  def scalaPBVersion = "0.7.4"
  override def scalaPBOptions = "flat_package,java_conversions"
}

Scoverage

This module allows you to generate code coverage reports for Scala projects with Scoverage via the scalac-scoverage-plugin.

To declare a module for which you want to generate coverage reports you can extends the mill.contrib.scoverage.ScoverageModule trait when defining your module. Additionally, you must define a submodule that extends the ScoverageTests trait that belongs to your instance of ScoverageModule.

import $ivy.`com.lihaoyi::mill-contrib-scoverage:$MILL_VERSION`
import mill.contrib.scoverage.ScoverageModule

object foo extends ScoverageModule  {
  def scalaVersion = "2.11.8"
  def scoverageVersion = "1.3.1"

  object test extends ScoverageTests {
    def ivyDeps = Agg(ivy"org.scalatest::scalatest:3.0.5")
    def testFrameworks = Seq("org.scalatest.tools.Framework")
  }
}

In addition to the normal tasks available to your Scala module, Scoverage modules introduce a few new tasks and changes the behavior of an existing one.

mill foo.scoverage.compile      # compiles your module with test instrumentation
                                # (you don't have to run this manually, running the test task will force its invocation)

mill foo.test                   # tests your project and collects metrics on code coverage
mill foo.scoverage.htmlReport   # uses the metrics collected by a previous test run to generate a coverage report in html format
mill foo.scoverage.xmlReport    # uses the metrics collected by a previous test run to generate a coverage report in xml format

The measurement data is available at out/foo/scoverage/data/, the html report is saved in out/foo/scoverage/htmlReport/, and the xml report is saved in out/foo/scoverage/xmlReport/.

TestNG

Provides support for TestNG.

To use TestNG as test framework, you need to add it to the TestModule.testFrameworks property.

// build.sc
import mill.scalalib._

object project extends ScalaModule {
  object test extends Tests{
    def testFrameworks = Seq("mill.testng.TestNGFramework")
  }
}

Tut

This module allows Tut to be used in Mill builds. Tut is a documentation tool which compiles and evaluates Scala code in documentation files and provides various options for configuring how the results will be displayed in the compiled documentation.

To declare a module that uses Tut you can extend the mill.contrib.tut.TutModule trait when defining your module.

This creates a Scala module which compiles markdown, HTML and .txt files in the tut folder of the module with Tut.

By default the resulting documents are simply placed in the Mill build output folder but they can be placed elsewhere by overriding the tutTargetDirectory task.

// build.sc

import $ivy.`com.lihaoyi::mill-contrib-tut:$MILL_VERSION`
import contrib.tut._

object example extends TutModule {
  def scalaVersion = "2.12.6"
  def tutVersion = "0.6.7"
}

This defines a project with the following layout:

build.sc
example/
    src/
    tut/
    resources/

In order to compile documentation we can execute the tut task in the module:

sh> mill example.tut

Configuration options

Twirl

Twirl templates support.

To declare a module that needs to compile twirl templates you must extend the mill.twirllib.TwirlModule trait when defining your module. Also note that twirl templates get compiled into scala code, so you also need to extend ScalaModule.

// build.sc
import mill.scalalib._

import $ivy.`com.lihaoyi::mill-contrib-twirllib:$MILL_VERSION`,  mill.twirllib._

object app extends ScalaModule with TwirlModule {
// ...
}

Twirl configuration options

Details

The following filesystem layout is expected:

build.sc
app/
    views/
        view1.scala.html
        view2.scala.html

TwirlModule adds the compileTwirl task to the module:

mill app.compileTwirl

(it will be automatically run whenever you compile your module)

This task will compile *.scala.html templates (and others, like *.scala.txt) into the out/app/compileTwirl/dest directory. This directory must be added to the generated sources of the module to be compiled and made accessible from the rest of the code:

// build.sc
import mill.scalalib._

import $ivy.`com.lihaoyi::mill-contrib-twirllib:$MILL_VERSION`,  mill.twirllib._

object app extends ScalaModule with TwirlModule {
  def twirlVersion = "1.3.15"
  def generatedSources = T{ Seq(compileTwirl().classes) }
}

To add additional imports to all of the twirl templates:

// build.sc
import mill.scalalib._

import $ivy.`com.lihaoyi::mill-contrib-twirllib:$MILL_VERSION`,  mill.twirllib._

object app extends ScalaModule with TwirlModule {
  def twirlVersion = "1.3.15"
  override def twirlAdditionalImports = Seq("my.additional.stuff._", "my.other.stuff._")
  def generatedSources = T{ Seq(compileTwirl().classes) }
}

as the result all templates will get this line at the top:

@import "my.additional.stuff._"
@import "my.other.stuff._"

Besides that, twirl compiler has default imports, at the moment these:

Seq(
    "_root_.play.twirl.api.TwirlFeatureImports._",
    "_root_.play.twirl.api.TwirlHelperImports._",
    "_root_.play.twirl.api.Html",
    "_root_.play.twirl.api.JavaScript",
    "_root_.play.twirl.api.Txt",
    "_root_.play.twirl.api.Xml"
)

These imports will always be added to every template. You don't need to list them if you override twirlAdditionalImports.

Example

There's an example project


About the Author: Haoyi is a software engineer, an early contributor to Scala.js, and the author of many open-source Scala tools such as Mill, the Ammonite REPL and FastParse.

If you've enjoy using Mill, or enjoyed using Haoyi's other open source libraries, please chip in (or get your Company to chip in!) via Patreon so he can continue his open-source work


Mill InternalsThirdparty Modules