Quoted Code

This doc page is specific to Scala 3, and may cover new concepts not available in Scala 2. Unless otherwise stated, all the code examples in this page assume you are using Scala 3.

Code blocks

A quoted code block '{ ... } is syntactically similar to a string quote " ... " with the difference that the first contains typed code. To insert code into other code, we can use the syntax $expr or ${ expr }, where expr is of type Expr[T]. Intuitively, the code directly within the quote ('{ ... }) is not executed now, while the code within the splice (${ ... }) is evaluated and the results spliced into the surrounding expression.

val msg = Expr("Hello")
val printHello = '{ print($msg) }
println(printHello.show) // print("Hello")

In general, the quote delays the execution while the splice makes it happen before the surrounding code. This generalisation allows us to also give meaning to a ${ ... } that is not within a quote. This evaluates the code within the splice at compile-time and places the result in the generated code. Due to some technical considerations, only top-level splices are allowed directly within inline definitions that we call a macro.

It is possible to write a quote within a quote, but this pattern is not common when writing macros.

Level consistency

One cannot simply write any arbitrary code within quotes and within splices, as one part of the program will live at compile-time and the other will live at runtime. Consider the following ill-constructed code:

def myBadCounter1(using Quotes): Expr[Int] = {
  var x = 0
  '{ x += 1; x }
}

The problem with this code is that x exists during compilation, but then we try to use it after the compiler has finished (maybe even in another machine). Clearly, it would be impossible to access its value and update it.

Now consider the dual version, where we define the variable at runtime and try to access it at compile-time:

def myBadCounter2(using Quotes): Expr[Int] = '{
  var x = 0
  ${ x += 1; 'x }
}

Clearly, this should not work as the variable does not exist yet.

To make sure you cannot write programs that contain these kinds of problems, we restrict the kinds of references allowed in quote environments.

We introduce levels as a count of the number of quotes minus the number of splices surrounding an expression or definition.

// level 0
'{ // level 1
  var x = 0
  ${ // level 0
    x += 1
    'x // level 1
  }
}

The system will allow references to global definitions such as println at any level, but will restrict references to local definitions. A local definition can only be accessed if it is defined at the same level as its reference. This will catch the errors in myBadCounter1 and myBadCounter2.

Even though we cannot refer to a variable inside of a quote, we can still pass its current value through a quote by lifting the value to an expression using Expr.apply.

Generics

When using type parameters or other kinds of abstract types with quoted code, we will need to keep track of some of these types explicitly. Scala uses erased-types semantics for its generics. This implies that types are removed from the program when compiling and the runtime does not have to track all types at runtime.

Consider the following code:

def evalAndUse[T](x: Expr[T])(using Quotes) = '{
  val x2: T = $x // error
  ... // use x2
}

Here, we will get an error telling us that we are missing a contextual Type[T]. Therefore, we can easily fix it by writing:

def evalAndUse[T](x: Expr[T])(using Type[T])(using Quotes) = '{
  val x2: T = $x
  ... // use x2
}

This code will be equivalent to this more verbose version:

def evalAndUse[T](x: Expr[T])(using t: Type[T])(using Quotes) = '{
  val x2: t.Underlying = $x
  ... // use x2
}

Note that Type has a type member called Underlying that refers to the type held within the Type; in this case, t.Underlying is T. Even if we use the Type implicitly, is generally better to keep it contextual as some changes inside the quote may require it. The less verbose version is usually the best way to write the types as it is much simpler to read. In some cases, we will not statically know the type within the Type and will need to use the t.Underlying to refer to it.

When do we need this extra Type parameter?

• When a type is abstract and it is used at a level that is higher than the current level.

When you add a Type contextual parameter to a method, you will either get it from another context parameter or implicitly with a call to Type.of:

evalAndUse(Expr(3))
// is equivalent to
evalAndUse[Int](Expr(3))(using Type.of[Int])

As you may have guessed, not every type can be used as a parameter to Type.of[..] out of the box. For example, we cannot recover abstract types that have already been erased:

def evalAndUse[T](x: Expr[T])(using Quotes) =
  given Type[T] = Type.of[T] // error
  '{
    val x2: T = $x
    ... // use x2
  }

But we can write more complex types that depend on these abstract types. For example, if we look for or explicitly construct a Type[List[T]], then the system will require a Type[T] in the current context to compile.

Good code should only add Types to the context parameters and never use them explicitly. However, explicit use is useful while debugging, though it comes at the cost of conciseness and clarity.

ToExpr

The Expr.apply method uses instances of ToExpr to generate an expression that will create a copy of the value.

object Expr:
  def apply[T](x: T)(using Quotes, ToExpr[T]): Expr[T] =
    summon[ToExpr[T]].apply(x)

ToExpr is defined as follows:

trait ToExpr[T]:
  def apply(x: T)(using Quotes): Expr[T]

The ToExpr.apply method will take a value T and generate code that will construct a copy of this value at runtime.

We can define our own ToExprs like:

given ToExpr[Boolean] with {
  def apply(x: Boolean)(using Quotes) =
    if x then '{true}
    else '{false}
}

given ToExpr[StringContext] with {
  def apply(stringContext: StringContext)(using Quotes) =
    val parts = Varargs(stringContext.parts.map(Expr(_)))
    '{ StringContext($parts*) }
}

The Varargs constructor just creates an Expr[Seq[T]] which we can efficiently splice as a varargs. In general, any sequence can be spliced with $mySeq* to splice it as a varargs.

Quoted patterns

Quotes can also be used to check if an expression is equivalent to another or to deconstruct an expression into its parts.

Matching exact expression

The simplest thing we can do is to check if an expression matches another known expression. Below, we show how we can match some expressions using case '{...} =>.

def valueOfBoolean(x: Expr[Boolean])(using Quotes): Option[Boolean] =
  x match
    case '{ true } => Some(true)
    case '{ false } => Some(false)
    case _ => None

def valueOfBooleanOption(x: Expr[Option[Boolean]])(using Quotes): Option[Option[Boolean]] =
  x match
    case '{ Some(true) } => Some(Some(true))
    case '{ Some(false) } => Some(Some(false))
    case '{ None } => Some(None)
    case _ => None

Matching partial expression

To make things more compact, we can also match a part of the expression using a splice ($) to match arbitrary code and extract it.

def valueOfBooleanOption(x: Expr[Option[Boolean]])(using Quotes): Option[Option[Boolean]] =
  x match
    case '{ Some($boolExpr) } => Some(valueOfBoolean(boolExpr))
    case '{ None } => Some(None)
    case _ => None

Matching types of expression

We can also match against code of an arbitrary type T. Below, we match against $x of type T and we get out an x of type Expr[T].

def exprOfOption[T: Type](x: Expr[Option[T]])(using Quotes): Option[Expr[T]] =
  x match
    case '{ Some($x) } => Some(x) // x: Expr[T]
    case '{ None } => Some(None)
    case _ => None

We can also check for the type of an expression:

def valueOf(x: Expr[Any])(using Quotes): Option[Any] =
  x match
    case '{ $x: Boolean } => valueOfBoolean(x) // x: Expr[Boolean]
    case '{ $x: Option[Boolean] }  => valueOfBooleanOption(x) // x: Expr[Option[Boolean]]
    case _ => None

Or similarly for a partial expression:

case '{ Some($x: Boolean) } => // x: Expr[Boolean]

Matching receiver of methods

When we want to match the receiver of a method, we need to explicitly state its type:

case '{ ($ls: List[Int]).sum } =>

If we would have written $ls.sum, we would not have been able to know the type of ls and which sum method we are calling.

Another common case where we need type annotations is for infix operations:

case '{ ($x: Int) + ($y: Int) } =>
case '{ ($x: Double) + ($y: Double) } =>
case ...

Matching function expressions

Coming soon

Matching types

So far, we assumed that the types within quote patterns would be statically known. Quote patterns also allow for type parameters, which we will see in this section.

Type parameters in patterns

Consider the function exprOfOption that we have already seen:

def exprOfOption[T: Type](x: Expr[Option[T]])(using Quotes): Option[Expr[T]] =
  x match
    case '{ Some($x: T) } => Some(x) // x: Expr[T]
                // ^^^ type ascription with type T
    ...

Note that this time we have added the T explicitly in the pattern, even though it could be inferred. By referring to the type parameter T in the pattern, we are required to have a given Type[T] in scope. This implies that $x: T will only match if x is of type Expr[T]. In this particular case, this condition will always be true.

Now consider the following variant where x is an optional value with a (statically) unknown element type:

def exprOfOptionOf[T: Type](x: Expr[Option[Any]])(using Quotes): Option[Expr[T]] =
  x match
    case '{ Some($x: T) } => Some(x) // x: Expr[T]
    case _ => None

This time, the pattern Some($x: T) will only match if the type of the Option is Some[T].

exprOfOptionOf[Int]('{ Some(3) })   // Some('{3})
exprOfOptionOf[Int]('{ Some("a") }) // None

Type variables in quoted patterns

Quoted code may contain types that are not known outside of the quote. We can match on them using pattern type variables. Just as in a normal pattern, the type variables are written using lower case names.

def exprOptionToList(x: Expr[Option[Any]])(using Quotes): Option[Expr[List[Any]]] =
  x match
    case '{ Some($x: t) } =>
                // ^^^ this binds the type `t` in the body of the case
      Some('{ List[t]($x) }) // x: Expr[List[t]]
    case '{ None } =>
      Some('{ Nil })
    case _ => None

The pattern $x: t will match an expression of any type and t will be bound to the type of the pattern. This type variable is only valid in the right-hand side of the case. In this example, we use it to construct the list List[t]($x) (List($x) would also work). As this is a type that is not statically, known we need a given Type[t] in scope. Luckily, the quoted pattern will automatically provide this for us.

The simple pattern case '{ $expr: tpe } => is very useful if we want to know the precise type of the expression.

val expr: Expr[Option[Int]] = ...
expr match
  case '{ $expr: tpe } =>
    Type.show[tpe] // could be: Option[Int], Some[Int], None, Option[1], Option[2], ...
    '{ val x: tpe = $expr; x } // binds the value without widening the type
    ...

In some cases we need to define a pattern variable that is referenced several times or has some type bounds. To achieve this, it is possible to create pattern variables at the start of the pattern using type t with a type pattern variable.

/**
 * Use: Converts a redundant `list.map(f).map(g)` to only use one call
 * to `map`: `list.map(y => g(f(y)))`.
 */
def fuseMap[T: Type](x: Expr[List[T]])(using Quotes): Expr[List[T]] = x match {
  case '{
    type u
    type v
    ($ls: List[`u`])
      .map($f: `u` => `v`)
      .map($g: `v` => T)
    } =>
    '{ $ls.map(y => $g($f(y))) }
  case _ => x
}

Here, we define two type variables u and v and then refer to them using `u` and `v`. We do not refer to them using u or v (without backticks) because those would be interpreted as new type variables with the same variable name. This notation follows the normal stable identifier patterns syntax. Furthermore, if the type variable needs to be constrained, we can add bounds directly on the type definition: case '{ type u <: AnyRef; ... } =>.

Note that the previous case could also be written as case '{ ($ls: List[u]).map[v]($f).map[T]($g) =>.

Quote types patterns

Types represented with Type[T] can be matched on using the patten case '[...] =>.

inline def mirrorFields[T]: List[String] = ${mirrorFieldsImpl[T]}

def mirrorFieldsImpl[T: Type](using Quotes): Expr[List[String]] =

  def rec[A : Type]: List[String] = Type.of[A] match
    case '[field *: fields] =>
      Type.show[field] :: rec[fields]
    case '[EmptyTuple] =>
      Nil
    case _ =>
      quotes.reflect.report.errorAndAbort("Expected known tuple but got: " + Type.show[A])

  Expr(rec)

mirrorFields[EmptyTuple]         // Nil
mirrorFields[(Int, String, Int)] // List("scala.Int", "java.lang.String", "scala.Int")
mirrorFields[Tuple]              // error: Expected known tuple but got: Tuple

As with expression quote patterns, type variables are represented using lower case names.

FromExpr

The Expr.value, Expr.valueOrAbort, and Expr.unapply methods uses intances of FromExpr to extract the value if possible.

extension [T](expr: Expr[T]):
  def value(using Quotes)(using fromExpr: FromExpr[T]): Option[T] =
    fromExpr.unapply(expr)

  def valueOrError(using Quotes)(using fromExpr: FromExpr[T]): T =
    fromExpr.unapply(expr).getOrElse(eport.throwError("...", expr))
end extension

object Expr:
  def unapply[T](expr: Expr[T])(using Quotes)(using fromExpr: FromExpr[T]): Option[T] =
    fromExpr.unapply(expr)

FromExpr is defined as follows:

trait FromExpr[T]:
  def unapply(x: Expr[T])(using Quotes): Option[T]

The FromExpr.unapply method will take a value x and generate code that will construct a copy of this value at runtime.

We can define our own FromExprs like so:

given FromExpr[Boolean] with {
  def unapply(x: Expr[Boolean])(using Quotes): Option[Boolean] =
    x match
      case '{ true } => Some(true)
      case '{ false } => Some(false)
      case _ => None
}

given FromExpr[StringContext] with {
  def unapply(x: Expr[StringContext])(using Quotes): Option[StringContext] = x match {
    case '{ new StringContext(${Varargs(Exprs(args))}*) } => Some(StringContext(args*))
    case '{     StringContext(${Varargs(Exprs(args))}*) } => Some(StringContext(args*))
    case _ => None
  }
}

Note that we handled two cases for StringContext. As it is a case class, it can be created with new StringContext or with StringContext.apply from the companion object. We also used the Varargs extractor to match the arguments of type Expr[Seq[String]] into a Seq[Expr[String]]. Then we used the Exprs to match known constants in the Seq[Expr[String]] to get a Seq[String].

The Quotes

The Quotes is the main entry point for the creation of all quotes. This context is usually just passed around through contextual abstractions (using and ?=>). Each quote scope will have its own Quotes. New scopes are introduced each time a splice is introduced (${ ... }). Though it looks like a splice takes an expression as argument, it actually takes a Quotes ?=> Expr[T]. Therefore, we could actually write it explicitly as ${ (using q) => ... }. This might be useful when debugging to avoid generated names for these scopes.

The method scala.quoted.quotes provides a simple way to use the current Quotes without naming it. It is usually imported along with the Quotes using import scala.quoted.*.

${ (using q1) => body(using q1) }
// equivalent to
${ body(using quotes) }

Warning: If you explicitly name a Quotes quotes, you will shadow this definition.

When we write a top-level splice in a macro, we are calling something similar to the following definition. This splice will provide the initial Quotes associated with the macro expansion.

def $[T](x: Quotes ?=> Expr[T]): T = ...

When we have a splice within a quote, the inner quote context will depend on the outer one. This link is represented using the Quotes.Nested type. Users of quotes will almost never need to use Quotes.Nested. These details are only useful for advanced macros that will inspect code and may encounter details of quotes and splices.

def f(using q1: Quotes) = '{
  ${ (using q2: q1.Nested) ?=>
      ...
  }
}

We can imagine that a nested splice is like the following method, where ctx is the context received by the surrounding quote.

def $[T](using q: Quotes)(x: q.Nested ?=> Expr[T]): T = ...

β-reduction

When we have a lambda applied to an argument in a quote '{ ((x: Int) => x + x)(y) }, we do not reduce it within the quote; the code is kept as-is. There is an optimisation that will β-reduce all lambdas directly applied to parameters to avoid the creation of a closure. This will not be visible from the quote’s perspective.

Sometimes it is useful to perform this β-reduction on the quotes directly. We provide the function Expr.betaReduce[T] that receives an Expr[T] and β-reduces if it contains a directly-applied lambda.

Expr.betaReduce('{ ((x: Int) => x + x)(y) }) // returns '{ val x = y; x + x }

Summon values

There are two ways to summon values in a macro. The first is to have a using parameter in the inline method that is passed explicitly to the macro implementation.

inline def setOf[T](using ord: Ordering[T]): Set[T] =
  ${ setOfCode[T]('ord) }

def setOfCode[T: Type](ord: Expr[Ordering[T]])(using Quotes): Expr[Set[T]] =
  '{ TreeSet.empty[T](using $ord) }

In this scenario, the context parameter is found before the macro is expanded. If not found, the macro will not be expanded.

The second way is using Expr.summon. This allows us to programatically search for distinct given expressions. The following example is similar to the previous example:

inline def setOf[T]: Set[T] =
  ${ setOfCode[T] }

def setOfCode[T: Type](using Quotes): Expr[Set[T]] =
  Expr.summon[Ordering[T]] match
    case Some(ord) => '{ TreeSet.empty[T](using $ord) }
    case _ => '{ HashSet.empty[T] }

The difference is that, in the second scenario, we expand the macro before the implicit search is performed. We can therefore write arbitrary code to handle the case when an Ordering[T] is not found. Here, we used HashSet instead of TreeSet because the former does not need an Ordering.

Quoted Type Classes

In the previous example we showed how to use the Expr[Ordering[T]] type class explicitly by leveraging the using argument clause. This is perfectly fine, but it is not very convenient if we need to use the type class multiple times. To show this we will use a powerCode function that can be used on any numeric type.

First, it can be useful to make Expr type class can make it a given parameter. To do this we do need to explicitly in power to powerCode because we have a given Numeric[Num] but require an Expr[Numeric[Num]]. But then we can ignore it in powerMacro and any other place that only passes it around.

inline def power[Num](x: Num, inline n: Int)(using num: Numeric[Num]) =
  ${ powerMacro('x, 'n)(using 'num) }

def powerMacro[Num: Type](x: Expr[Num], n: Expr[Int])(using Expr[Numeric[Num]])(using Quotes): Expr[Num] =
  powerCode(x, n.valueOrAbort)

To use a this type class we need a given Numeric[Num] but we have a Expr[Numeric[Num]] and therefore we need to splice this expression in the generated code. To make it available we can just splice it in a given definition.

def powerCode[Num: Type](x: Expr[Num], n: Int)(using num: Expr[Numeric[Num]])(using Quotes): Expr[Num] =
  if (n == 0) '{ $num.one }
  else if (n % 2 == 0) '{
    given Numeric[Num] = $num
    val y = $x * $x
    ${ powerCode('y, n / 2) }
  }
  else '{
    given Numeric[Num] = $num
    $x * ${ powerCode(x, n - 1) }
  }