sbt has moved

sbt has now completely moved to GitHub.

See https://github.com/harrah/xsbt/wiki/Input-Tasks.

Method Tasks

As shown in several examples on the Triggered Execution page, a method task accepts an array of arguments and creates the task to be run from those arguments. Usage of a method task from the interactive prompt looks like:

> run 10000 5 0.3

or

> test-only example.TestSpecification

Defining your own method task is similar to defining a normal action, as shown in the following example.

lazy val example =
   task { args =>
     if(args.length == 2)
       actionConstructor(args(0).toInt, args(1))
     else
       task { Some("Usage: example <integer> <string>") }
   }
def actionConstructor(a: Int, b: String) =
    task {
      println("Arguments were: " + a + " and " + b)
      None
    }

The type of args is Array[String].

You can then call example like:

> example 5 text
Arguments were: 5 and text

Additionally, you can use quotes to include whitespace in arguments. Quotes must be escaped (\") and so must backslashes (\\):

> example "5" "Text with \"spaces\" and a backslash \\ included."
Arguments were: 5 and Text with "spaces" and a backslash \ included.

You can provide strings for basic tab completion for method tasks using the completeWith method.

  def testNames: Seq[String] = ...

  lazy val myTestOnly =
    task { args => ... } completeWith( testNames )

The argument to completeWith is call-by-name and is re-evaluated after a command is run.

In a multi-project setting, method tasks can only be run from the project they are defined on.