scala: tracing implicits selection and other code magics

问题

When trying to figure how a library works, implicit conversions are confusing. For example, looking at an expression like 'val foo: Foo = 1', what converts 1 to Foo?

Is it possible to instruct the scala library (or REPL) to print out the code paths that are executing while evaluating an expression?

回答1:

You can add "-Xprint:typer" to the compiler command line (or "-Ybrowse:typer" for a swing GUI browser) to see the code with the conversions explicitly applied.

回答2:

As an alternative to printing out the conversions, one must realize implicits can't just come out of the blue. You have to bring them into scope in some way. The alternatives are:

1. Explicit import statement. Watch out for import x.y._ when y is an object, as this is the only way to bring in an implicit into scope.
2. The object companion of the class that is being converted into something else.
3. The object companion of the target class, as long as that target is made explicit somehow (such as in your example).

Note that the object scala.Predef is all imported into scope by default, which is how Scala's default implicits get into scope.

回答3:

scalac -print prints the code after implicit type conversions where applied.

class A{
    val x : String = "hi" drop 1 
}

Will result in:

package <empty> {
  class A extends java.lang.Object with ScalaObject {
    @remote def $tag(): Int = scala.ScalaObject$class.$tag(A.this);
    private[this] val x: java.lang.String = _;
    <stable> <accessor> def x(): java.lang.String = A.this.x;
    def this(): A = {
      A.super.this();
      A.this.x = scala.this.Predef.forceRandomAccessCharSeq(
        scala.this.Predef.stringWrapper("hi").drop(1));
      ()
    }
  }
}

来源：https://stackoverflow.com/questions/1818181/scala-tracing-implicits-selection-and-other-code-magics