从Scala宏注释获取参数

时间:2015-09-17 13:21:18

标签: scala annotations scala-macros

所以我在函数上有一个注释(DefDef)。此注释具有参数。 但是,我对如何从构造函数中获取参数感到困惑。

用法示例:

class TestMacro {
  @Foo(true)
  def foo(): String = ""
  foo
}

以下是注释的代码:

class Foo(b: Boolean) extends StaticAnnotation {
  def macroTransform(annottees: Any*) = macro Foo.impl
}

object Foo {
  def impl(c: whitebox.Context)(annottees: c.Tree*): c.Expr[Any] = {
    import c.universe._
    //how do I get value of `b` here???
    c.abort(c.enclosingPosition, "message")
  }
}

2 个答案:

答案 0 :(得分:9)

这个怎么样:

val b: Boolean = c.prefix.tree match {
    case q"new Foo($b)" => c.eval[Boolean](c.Expr(b))
}

为了完整起见,这是完整的来源:

import scala.reflect.macros.Context
import scala.language.experimental.macros
import scala.annotation.StaticAnnotation
import scala.annotation.compileTimeOnly
import scala.reflect.api.Trees
import scala.reflect.runtime.universe._

class Foo(b: Boolean) extends StaticAnnotation {
  def macroTransform(annottees: Any*) :Any = macro FooMacro.impl
}

object FooMacro {
  def impl(c: Context)(annottees: c.Expr[Any]*): c.Expr[Any] = {
    import c.universe._
    val b: Boolean = c.prefix.tree match {
        case q"new Foo($b)" => c.eval[Boolean](c.Expr(b))
    }
    c.abort(c.enclosingPosition, "message")
  }
}

答案 1 :(得分:2)

如果您想使用具有可选命名参数的静态注释,这个答案会显示Federico技术的变体。在这种情况下,您需要考虑大小写匹配语句中可能的调用表达式。可以显式命名可选参数,可以在没有名称的情况下给出,也可以不存在。其中每个都在编译时显示为c.prefix.tree中的单独模式,如下所示。

@compileTimeOnly("Must enable the Scala macro paradise compiler plugin to expand static annotations")
class noop(arg1: Int, arg2: Int = 0) extends StaticAnnotation {
  def macroTransform(annottees: Any*): Any = macro AnnotationMacros.noop
}

class AnnotationMacros(val c: whitebox.Context) {
  import c.universe._

  // an annotation that doesn't do anything:
  def noop(annottees: c.Expr[Any]*): c.Expr[Any] = {
    // cases for handling optional arguments
    val (arg1q, arg2q) = c.prefix.tree match {
      case q"new noop($arg1, arg2 = $arg2)" => (arg1, arg2)  // user gave named arg2
      case q"new noop($arg1, $arg2)" => (arg1, arg2)         // arg2 without name
      case q"new noop($arg1)" => (arg1, q"0")                // arg2 defaulted
      case _ => c.abort(c.enclosingPosition, "unexpected annotation pattern!")
    }

    // print out the values
    println(s"arg1= ${evalTree[Int](arg1q)}   arg2= ${evalTree[Int](arg2q)}")

    // just return the original annotee:
    annottees.length match {
      case 1 => c.Expr(q"{ ${annottees(0)} }")
      case _ => c.abort(c.enclosingPosition, "Only one annottee!")
    }
  }

  def evalTree[T](tree: Tree) = c.eval(c.Expr[T(c.untypecheck(tree.duplicate)))
}

以下是一个名为arg2的示例调用,因此它将匹配第一个模式 - case q"new noop($arg1, arg2 = $arg2)" - 上面:

object demo {
  // I will match this pattern: case q"new noop($arg1, arg2 = $arg2)"
  @noop(1, arg2 = 2)
  trait someDeclarationToAnnotate
}

另请注意,由于这些模式的工作方式,您必须在宏代码中明确提供默认参数值,遗憾的是有点hacky,但最终评估的类不可用。

作为一个实验,我尝试通过调用evalTree[scope.of.class.noop](c.prefix.tree)来实际创建类,但Scala编译器抛出一个错误,因为它认为注释宏代码中的注释引用是非法的。