Simple Type Inference in Scala - scala

I have been looking at type inference in Scala and there are a couple of things I'd like to understand a bit better around why expression/method-return types have to be explicitly declared in a few cases.
Explicit return declaration
Example (works if return keyword is ommitted):
def upCase(s: String) = {
if (s.length == 0)
return s // COMPILE ERROR - forces return type of upCase to be declared.
else
s.toUpperCase()
}
Why can't I use the explicitly typed parameter as a return value without declaring the return type? And that's not only for direct parameter references, just for any 'type-inferable' expression.
Method overloading
Example (fails to compile when the second joiner method is added):
def joiner(ss: List[String], sep: String) = ss.mkString(sep)
def joiner(ss: List[String]) = joiner(strings, " ") // COMPILE ERROR WHEN ADDED

Well most obvious answer is: because it stated in specification see part 6.20 of scala reference. But why it was designed this way is indeed very intresting question. I suspect it connected to the fact that compiler can't predict that expression will be the last one, since return changes execution flow.
EDIT:
Consider if return doesn't require explicit return type following code:
def bar() = {
if(guard())
return "SS"
else if(gurard1())
return true
2
}
that return type should bar have in this situation? Well there is option with most common supertype, but I think it will get us to returning Any in many cases. Well this is just my thoughts which may be totally incorrect =)

The type of a function or method is inferred from the type of its last statement. Usually, that's an expression.
Now, "return" breaks the control flow. It is an "immediate interrupt", so to speak. Because of that, the normal rules used to infer the type of an expression can't be used anymore. It still could be done, of course, but I'm guessing the cost in compiler complexity was deemed to high for the return.
Here's an example of how the flow is broken:
def toNumber(s: String) = {
if (s == null)
return ""
if (s matches """\d+""")
s.toInt
else
0
}
Normally, the type of the second if statement would be used to infer the type of the whole function. But the return on the first if introduces a second return point from the function, so this rule won't work.

Type inference infers the return type of a method when it can, which is more or less in any case that the method isn't recursive.
Your example would work if you changed it to:
def upCase(s: String) = {
if (s.length == 0)
s // note: no return
else
s.toUpperCase()
}
I don't know offhand why the return changes this.

Disclaimer - this answer was directed to the question as it was originally posted
Scala's type inference already does infer the return type of a method / expression:
scala> def foo(s : String) = s + " Hello"
foo: (String)java.lang.String
scala> var t = foo("World")
t: java.lang.String = World Hello
and:
scala> def bar( s : String) = s.toInt
bar: (String)Int
scala> var i = bar("3")
i: Int = 3
and:
scala> var j = if (System.getProperty("user.name") == "oxbow") 4 else "5".toInt
j: Int = 5
EDIT - I didn't realize that the inclusion of the return keyword meant that the return type of an expression had to be explicitly declared: I've pretty much stopped using return myself - but it's an interesting question. For the joiner example, the return type must be declared because of overloading. Again, I don't know the details as to why and would be interested to find out. I suspect a better-phrased question subject would elicit an answer from the likes of James Iry, Dan Spiewak or Daniel Sobral.

I suspect the method overloading (lack of) inference is related to the similar problem with recursive calls, because if the overloaded methods doesn't call each other, it works perfectly:
def joiner1(ss: List[String], sep: String) = ss.mkString(sep)
def joiner(ss: List[String], sep: String) = ss.mkString(sep)
def joiner(ss: List[String]) = joiner1(ss, " ")
There's two overloaded joiner methods, but the types are inferred correctly the code compiles.

Related

Scala - Function's implicit return type

I am new to scala, and got a little doubt about function definition & default return type.
Here is a function definition:
def wol(s: String) = s.length.toString.length
The prompt says it's:
wol: (s: String)Int
But, the code didn't specify return type explicitly, shouldn't it default to Unit, which means void in Java.
So, what is the rules for default return type of a Scala function?
The return type in a function is actually the return type of the last expression that occurs in the function. In this case it's an Int, because #length returns an Int.
This is the work done by the compiler when it tries to infer the type. If you don't specify a type, it automatically gets inferred, but it's not necessarily Unit. You could force it to be that be stating it:
def wol(s: String): Unit = s.length.toString.length
EDIT [syntactic sugar sample]
I just remembered something that might be connected to your previous beliefs. When you define a method without specifying its return type and without putting the = sign, the compiler will force the return type to be Unit.
def wol(s: String) {
s.length.toString.length
}
val x = wol("") // x has type Unit!
IntelliJ actually warns you and gives the hint Useless expression. Behind the scene, the #wol function here is converted into something like:
// This is actually the same as the first function
def wol(s: String): Unit = { s.length.toString.length }
Anyway, as a best practice try to avoid using this syntax and always opt for putting that = sign. Furthermore if you define public methods try to always specify the return type.
Hope that helps :)

Any class in scala

I am beginner in scala language, I am confused with Any class in scala.
def f(x: Any) = println(x)
is above code represents that x variable can be of any datatype(example:int,string etc.)
rewritten code:
def f(x:Any)=x+5
<console>:13: error: type mismatch;
found : Int(5)
required: String
def f(x:Any)=x+5
if x can accept any type then why am I getting above error. I might confused understanding of any in scala. Please correct me.
In a statically typed language you can only call a method m on a value x of type A if m is defined by A. By the nature of Any, there aren't any useful methods on Any that you could call (except a few things like toString or hashCode), certainly no plus operation is defined. Imagine you passed a Boolean into that method, which is allowed since Boolean is a sub-type of Any. If the compiler allowed your code, it would run into a problem, because there is no such thing as + on a Boolean. In a dynamically typed language you could run that code and would then encounter a runtime error.
The error message looks weird, because you can concatenate strings with + and due to an implicit conversion it is possible to concatenate things with strings:
def f(x: Any) = x + "hello" // implicitly converts x to a string
f(true) // "truehello"
This is a source of great confusion and hopefully will disappear from the language. If you used a different method, the error would be more obvious:
def f(x:Any)=x-5
<console>:54: error: value - is not a member of Any
def f(x:Any)=x-5
^
If you look at println() sources you will see:
public void println(Object x) {
String s = String.valueOf(x);
synchronized (this) {
print(s);
newLine();
}
}
where String.valueOf is
public static String valueOf(Object obj) {
return (obj == null) ? "null" : obj.toString();
}
Class Any already has toString method but doesn't have + method (you know that all operators such +, - etc. are methods, see http://tomjefferys.blogspot.com/2011/11/operator-overloading-in-scala.html).

What's the difference between using and no using a "=" in Scala defs ?

What the difference between the two defs below
def someFun(x:String) { x.length }
AND
def someFun(x:String) = { x.length }
As others already pointed out, the former is a syntactic shortcut for
def someFun(x:String): Unit = { x.length }
Meaning that the value of x.length is discarded and the function returns Unit (or () if you prefer) instead.
I'd like to stress out that this is deprecated since Oct 29, 2013 (https://github.com/scala/scala/pull/3076/), but the warning only shows up if you compile with the -Xfuture flag.
scala -Xfuture -deprecation
scala> def foo {}
<console>:1: warning: Procedure syntax is deprecated. Convert procedure `foo` to method by adding `: Unit =`.
def foo {}
foo: Unit
So you should never use the so-called procedure syntax.
Martin Odersky itself pointed this out in his Scala Day 2013 Keynote and it has been discussed in the scala mailing list.
The syntax is very inconsistent and it's very common for a beginner to hit this issue when learning the language. For this reasons it's very like that it will be removed from the language at some point.
Without the equals it is implicitly typed to return Unit (or "void"): the result of the body is fixed - not inferred - and any would-be return value is discarded.
That is, def someFun(x:String) { x.length } is equivalent to def someFun(x:String): Unit = { x.length }, neither of which are very useful here because the function causes no side-effects and returns no value.
The "equals form" without the explicit Unit (or other type) has the return type inferred; in this case that would be def someFun(x:String): Int = { x.length } which is more useful, albeit not very exciting.
I prefer to specify the return type for all exposed members, which helps to ensure API/contract stability and arguably adds clarity. For "void" methods this is trivial done by using the Procedure form, without the equals, although it is a stylistic debate of which is better - opponents might argue that having the different forms leads to needless questions about such ;-)
The former is
def someFun(x: String): Unit = {
x.length
() // return unit
}
And the latter is
def someFun(x: String): Int = {
x.length // returned
}
Note that the Scala Style guide always recommend using '=', both in
method declaration
Methods should be declared according to the following pattern:
def foo(bar: Baz): Bin = expr
function declaration
Function types should be declared with a space between the parameter type, the arrow and the return type:
def foo(f: Int => String) = ...
def bar(f: (Boolean, Double) => List[String]) = ...
As per Scala-2.10, using equals sign is preferred. Infact you must use equals sign in call declarations except the definitions returning Unit.
As such there is no difference but the first one is not recommended anymore and it should not be used.

Choice of implicit conversion location with a block

In the following code, the implicit conversion is applied around the println(2) line; I had foolishly expected it to apply around the entire block { println(1); println(2) }. How should I reason about where the compiler places the implicit?
object Executor {
private var runnable: Runnable = _
def setRunnable(runnable: Runnable) {
this.runnable = runnable
}
def execute() { runnable.run() }
}
object Run extends App {
implicit def blockToRunnable(p: ⇒ Any): Runnable =
new Runnable { def run() = p }
Executor.setRunnable {
println(1)
println(2)
}
println("Before execute")
Executor.execute()
}
I rationalize this behavior like this: according to the spec, the type of a block {s1; s2; ...; sn; e } is the type of the last expression e.
So the compiler takes e and type checks it against Runnable. That fails, so it searches for an implicit conversion that will convert e to Runnable. So it would like this:
{ s1; s2; ... sn; convert(e) }
This is confirmed with scala -Xprint:typer on this small example:
class A
implicit def convert(a: A): String = a.toString
def f(s: String) { println(s) }
f{ println(1); new A }
prints:
private[this] val res0: Unit = $line3.$read.$iw.$iw.f({
scala.this.Predef.println(1);
$line2.$read.$iw.$iw.convert(new $line1.$read.$iw.$iw.A())
});
According to the spec, an implicit conversion is applied when the type of an expression does not match the expected type. The key observation is how the expected type is threaded when typing blocks.
if an expression e is of type T, and T does not conform to the expression’s expected type pt. In this case an implicit v is searched which is applicable to e and whose result type conforms to pt.
In Section 6.11 Blocks, the expected type of a block's last expression is defined as
The expected type of the final expression e is the expected type of the block.
Given this spec, it seems the compiler has to behave this way. The expected type of the block is Runnable, and the expected type of println(2) becomes Runnable as well.
A suggestion: if you want to know what implicits are applied, you can use a nightly build for 2.1 of the Scala IDE for Eclipse. It can 'highlight implicits'.
Edited: I admit it is surprising when there's a call-by-name implicit in scope.
The problem is that you are thinking of blocks as if they were thunks, as if they were pieces of code. They aren't. { a; b; c } is not a piece of code that can be passed around.
So, how should you reason about implicits? Actually, how should you reason about views, which are implicit conversions. Views are applied to the value that needs to be changed. In your example, the value of
{
println(1)
println(2)
}
is being passed to setRunnable. The value of a block is the value of its last expression, so it is passing the result of println(2) to setRunnable. Since that is Unit and setRunnable requires a Runnable, then an implicit is searched for and found, so println(2) is passed to the grossly misnamed blockToRunnable.
Bottom line is, and this is an advice I have given many times already on Stack Overflow (lots of people try to do the same thing) is to get the following in your head:
THERE ARE NO BLOCKS IN SCALA.
There are functions, but not blocks.
Technically, that statement is incorrect -- there are blocks in Scala, but they are not what you think they are, so just completely remove them from your mind. You can learn what blocks in Scala are latter, from a clean slate. Otherwise, you're bound to try to get them to work in ways they don't, or infer that things work in a certain way when they work in a different way.
I liked a lot the explanation given in the first scala puzzle.
In other words, what will be the output of:
List(1, 2).map { i => println("Hi"); i + 1 }
List(1, 2).map { println("Hi"); _ + 1 }

When is a return type required for methods in Scala?

The Scala compiler can often infer return types for methods, but there are some circumstances where it's required to specify the return type. Recursive methods, for example, require a return type to be specified.
I notice that sometimes I get the error message "overloaded method (methodname) requires return type", but it's not a general rule that return types must always be specified for overloaded methods (I have examples where I don't get this error).
When exactly is it required to specify a return type, for methods in general and specifically for overloaded methods?
The Chapter 2. Type Less, Do More of the Programming Scala book mentions:
When Explicit Type Annotations Are Required.
In practical terms, you have to provide explicit type annotations for the following situations:
Method return values in the following cases:
When you explicitly call return in a method (even at the end).
When a method is recursive.
When a method is overloaded and one of the methods calls another. The calling method needs a return type annotation.
When the inferred return type would be more general than you intended, e.g., Any.
Example:
// code-examples/TypeLessDoMore/method-nested-return-script.scala
// ERROR: Won't compile until you put a String return type on upCase.
def upCase(s: String) = {
if (s.length == 0)
return s // ERROR - forces return type of upCase to be declared.
else
s.toUpperCase()
}
Overloaded methods can sometimes require an explicit return type. When one such method calls another, we have to add a return type to the one doing the calling, as in this example.
// code-examples/TypeLessDoMore/method-overloaded-return-script.scala
// Version 1 of "StringUtil" (with a compilation error).
// ERROR: Won't compile: needs a String return type on the second "joiner".
object StringUtil {
def joiner(strings: List[String], separator: String): String =
strings.mkString(separator)
def joiner(strings: List[String]) = joiner(strings, " ") // ERROR
}
import StringUtil._ // Import the joiner methods.
println( joiner(List("Programming", "Scala")) )
The two joiner methods concatenate a List of strings together.
The first method also takes an argument for the separator string.
The second method calls the first with a “default” separator of a single space.
If you run this script, you get the following error.
... 9: error: overloaded method joiner needs result type
def joiner(strings: List[String]) = joiner(strings, "")
Since the second joiner method calls the first, it requires an explicit String return type. It should look like this:
def joiner(strings: List[String]): String = joiner(strings, " ")
Basically, specifying the return type can be a good practice even though Scala can infer it.
Randall Schulz comments:
As a matter of (my personal) style, I give explicit return types for all but the most simple methods (basically, one-liners with no conditional logic).
Keep in mind that if you let the compiler infer a method's result type, it may well be more specific than you want. (E.g., HashMap instead of Map.)
And since you may want to expose the minimal interface in your return type (see for instance this SO question), this kind of inference might get in the way.
And about the last scenario ("When the inferred return type would be more general than you intended"), Ken Bloom adds:
specify the return type when you want the compiler to verify that code in the function returns the type you expected
(The faulty code which triggers a "more general than expected return type was:
// code-examples/TypeLessDoMore/method-broad-inference-return-script.scala
// ERROR: Won't compile. Method actually returns List[Any], which is too "broad".
def makeList(strings: String*) = {
if (strings.length == 0)
List(0) // #1
else
strings.toList
}
val list: List[String] = makeList() // ERROR
, which I incorrectly interpreted and List[Any] because returning an empty List, but Ken called it out:
List(0) doesn't create a list with 0 elements.
It creates a List[Int] containing one element (the value 0).
Thus a List[Int] on one conditional branch and a List[String] on the other conditional branch generalize to List[Any].
In this case, the typer isn't being overly-general -- it's a bug in the code.
)