Why were the case classes without a parameter list deprecated from Scala? And why does compiler suggest to use ()
as parameter list instead?
EDIT :
Someone please answer my second question... :|
Why were the case classes without a parameter list deprecated from Scala? And why does compiler suggest to use ()
as parameter list instead?
EDIT :
Someone please answer my second question... :|
Without parameters, every instance of the case class is indistinguishable and hence is essentially a constant. Use an object for that case.
It is really easy to accidentally use a no-arg case class incorrectly as a pattern.
scala> case class Foo
warning: there were deprecation warnings; re-run with -deprecation for details
defined class Foo
scala> (new Foo: Any) match { case Foo => true; case _ => false }
res10: Boolean = false
Instead of:
scala> (new Foo: Any) match { case _: Foo => true; case _ => false }
res11: Boolean = true
Or better:
scala> case object Bar
defined module Bar
scala> (Bar: Any) match { case Bar => true; case _ => false }
res12: Boolean = true
UPDATE Hopefully the transcript below will demonstrate why an empty parameter list is preferred to the deprecated missing parameter list.
scala> case class Foo() // Using an empty parameter list rather than zero parameter lists.
defined class Foo
scala> Foo // Access the companion object Foo
res0: Foo.type = <function0>
scala> Foo() // Call Foo.apply() to construct an instance of class Foo
res1: Foo = Foo()
scala> case class Bar
warning: there were deprecation warnings; re-run with -deprecation for details
defined class Bar
scala> Bar // You may expect this to construct a new instance of class Bar, but instead
// it references the companion object Bar
res2: Bar.type = <function0>
scala> Bar() // This calls Bar.apply(), but is not symmetrical with the class definition.
res3: Bar = Bar()
scala> Bar.apply // Another way to call Bar.apply
res4: Bar = Bar()
A case object would usually still be preferred over an empty parameter list.