Need help figuring out scala compiler errors - scala

I have been working on a project in scala, but I am getting some error messages that I don't quite understand. The classes that I am working with are relatively simple.
For example:
abstract class Shape
case class Point(x: Int, y: Int) extends Shape
case class Polygon(points: Point*) extends Shape
Now suppose that I create a Polygon:
val poly = new Polygon(new Point(2,5), new Point(7,0), new Point(3,1))
Then if I attempt to determine the location and size of the smallest possible rectangle that could contain the polygon, I get various errors that I don't quite understand.
Below are snippets of different attempts and the corresponding error messages that they produce.
val upperLeftX = poly.points.reduceLeft(Math.min(_.x, _.x))
Gives the error:
"missing parameter type for expanded function ((x$1) => x$1.x)"
val upperLeftX =
poly.points.reduceLeft((a: Point, b: Point) => (Math.min(a.x, b.x)))
Gives this error:
"type mismatch;
found : (Point, Point) => Int
required: (Any, Point) => Any"
I am very confused about both of these error messages. If anyone could explain more clearly what I am doing incorrectly, I would really appreciate it. Yes, I see that the second error says that I need type "Any" but I don't understand exactly how to implement a change that would work as I need it. Obviously simply changing "a: Point" to "a: Any" is not a viable solution, so what am I missing?

The type of reduceLeft is reduceLeft[B >: A](op: (B, A) => B): B, A is Point, and you are trying to apply it to (a: Point, b: Point) => (Math.min(a.x, b.x)).
The compiler reasons thus: Math.min(a.x, b.x) returns Int, so Int must be a subtype of B. And B must also be a supertype of Point. Why? B is the type of the accumulator, and its initial value is the first Point in your Polygon. That's the meaning of B >: A.
The only supertype of Int and Point is Any; so B is Any and the type of op should be (Any, Point) => Any, just as the error message says.

This is Scala 2.8.0.RC2
scala> abstract class Shape
defined class Shape
scala> case class Point(x: Int, y: Int) extends Shape
defined class Point
scala> case class Polygon(points: Point*) extends Shape
defined class Polygon
scala> val poly = new Polygon(new Point(2,5), new Point(7,0), new Point(3,1))
poly: Polygon = Polygon(WrappedArray(Point(2,5), Point(7,0), Point(3,1)))
scala> val upperLeftX = poly.points.reduceLeft((a:Point,b:Point) => if (a.x < b.x) a else b)
upperLeftX: Point = Point(2,5)
reduceLeft requires here a function of the type (Point, Point) => Point. (more precisely (B, Point) => B with B with a lower bound to Point. See Scaladoc at the method reduceLeft.

Another alternative is poly.points.foldLeft(Int.MaxValue)((b, a) => Math.min(b, a.x)), which should also work with Scala 2.7.x. The differences compared to the reduceLeft version are
you have a start value (Int.MaxValue in our case, any real data will be smaller or equal to this)
there are no constraints between the type of the elements and the type of the result, like the lower bound constraint for reduceLeft
Nevertheless Eastsun's solution is more elegant.
BTW if you already have case classes you can ommit the new keyword, and can use the automatically generated factory method in the companion object. So the line creating the poly becomes val poly = Polygon(Point(2,5), Point(7,0), Point(3,1)), which is a bit easier to read.

I see everyone seems to have latched onto the second snippet, so I'll answer the first one:
val upperLeftX = poly.points.reduceLeft(Math.min(_.x, _.x))
You intended that to mean this:
val upperLeftX = poly.points.reduceLeft((a, b) => Math.min(a.x, b.x))
However, that's not how underscore works. There are many meanings to underscore, but two of them are relevant here.
First, it may mean a partial function application. For example, Math.min(_, 0) would partially apply the parameters to min, and return a function that apply the remaining ones. In other words, it is equivalent to x => Math.min(x, 0), ignoring the type annotations. At any rate, this meaning only applies if the underscore is all by itself in the place of one (or more) of the parameters.
That, however, is not the case in your example, because you added a .x after the underscore. If the underscore appears in any kind of expression, such as the method call in your example, then that underscore is a placeholder for a parameter in an anonymous function.
In this second meaning it is particularly important to understand the boundaries of the anonymous function. Specifically, the anonymous function will be delimited by the innermost parenthesis or curly brackets that encloses it, or by any comma.
Now, applying that rule to the expression in the first snippet means that snipper is seen by the compiler like this:
val upperLeftX = poly.points.reduceLeft(Math.min(a => a.x, b => b.x))
So, there are two problems here. First, you are passing two functions to min instead of two doubles. Second, because min isn't expecting to receive functions, the compiler can't infer what the type of these functions might be. Since you did not provide any information about the types of a and b above, it complains about that.
If you did provide such types, the error message would be something like this:
<console>:6: error: type mismatch;
found : Int
required: ?{val x: ?}

Related

Scala - pattern matching traits that have "self type" as type argument gives unchecked warning

I'm trying to define "immutable setter traits", and generic functions for those.
I have a working implementation, but i'm bit disturbed about the "unchecked" warnings from the pattern matching. I'm not really sure what i can do about it.
type Point = (Double, Double)
trait Sizable[A] {
this: A =>
def size: Point
/* immutable object value setter, returns a new of the same object*/
def size(point: Point): A with Sizable[A]
}
def partitionToSizable[T](elements: List[T]):
(List[T], List[T with Sizable[T]]) =
elements.foldLeft((List[T](), List[T with Sizable[T]]()))((p, c) =>
c match {
case a: T with Sizable[T] => (p._1, p._2 ++ List(a))
case a => (p._1 ++ List(a), p._2)
})
The code above demonstrates the problem.
I'm not even sure how big of an issue is that T being unchecked, since all the elements in the list will have a type of T, and the point of the pattern matching is not to determine if it's type is T since we already know that.
In theory Sizable will always have type of T because of the signature of it's enclosing function.
If it's there is no other solution i'd at least like to suppress the warning. #unchecked annotations does not seem to suppress the warning.
If i modify case a: T with Sizable[T] to case a: Sizable[_] it will not compile, since the result type will obviously not confirm to T.
ClassTags or TypeTags might solve the warning, but i suspect they are not necessary really. (also that might have a performance overhead and TypeTags don't work with Scala.js)
I think just case a: Sizable[T #unchecked] => a.size((15,20)) should work.
The main problem here is in partitionToSizable. It's taking a List[T] and then trying to partition it into a List[T] and a List[T with Sizable[T]]. You are taking a list of a single type and then saying that one type T is actually two different types T and T with Sizable[T]. This indicates an issue with your type design. I would recommend solving that first.
One solution might be to recognise that things that are sizable and things that are not sizable should be represented as two different types. Then you can use the following signature:
def partitionToSizable[T1, T2 <: Sizable[T2]](
elements: List[Either[T1, T2]]): (List[T1], List[T2])
Then you don't need to cast; you can pattern match or fold on the Either.

Scala: Typecast without explicitly known type parameter

Consider the following example:
case class C[T](x:T) {
def f(t:T) = println(t)
type ValueType = T
}
val list = List(1 -> C(2), "hello" -> C("goodbye"))
for ((a,b) <- list) {
b.f(a)
}
In this example, I know (runtime guarantee) that the type of a will be some T, and b will have type C[T] with the same T. Of course, the compiler cannot know that, hence we get a typing error in b.f(a).
To tell the compiler that this invocation is OK, we need to do a typecast à la b.f(a.asInstanceOf[T]). Unfortunately, T is not known here. So my question is: How do I rewrite b.f(a) in order to make this code compile?
I am looking for a solution that does not involve complex constructions (to keep the code readable), and that is "clean" in the sense that we should not rely on code erasure to make it work (see the first approach below).
I have some working approaches, but I find them unsatisfactory for various reasons.
Approaches I tried:
b.asInstanceOf[C[Any]].f(a)
This works, and is reasonably readable, but it is based on a "lie". b is not of type C[Any], and the only reason we do not get a runtime error is because we rely on the limitations of the JVM (type erasure). I think it is good style only to use x.asInstanceOf[X] when we know that x is really of type X.
b.f(a.asInstanceOf[b.ValueType])
This should work according to my understanding of the type system. I have added the member ValueType to the class C in order to be able to explicitly refer to the type parameter T. However, in this approach we get a mysterious error message:
Error:(9, 22) type mismatch;
found : b.ValueType
(which expands to) _1
required: _1
b.f(a.asInstanceOf[b.ValueType])
^
Why? It seems to complain that we expect type _1 but got type _1! (But even if this approach works, it is limited to the cases where we have the possibility to add a member ValueType to C. If C is some existing library class, we cannot do that either.)
for ((a,b) <- list.asInstanceOf[List[(T,C[T]) forSome {type T}]]) {
b.f(a)
}
This one works, and is semantically correct (i.e., we do not "lie" when invoking asInstanceOf). The limitation is that this is somewhat unreadable. Also, it is somewhat specific to the present situation: if a,b do not come from the same iterator, then where can we apply this type cast? (This code also has the side effect of being too complex for Intelli/J IDEA 2016.2 which highlights it as an error in the editor.)
val (a2,b2) = (a,b).asInstanceOf[(T,C[T]) forSome {type T}]
b2.f(a2)
I would have expected this one to work since a2,b2 now should have types T and C[T] for the same existential T. But we get a compile error:
Error:(10, 9) type mismatch;
found : a2.type (with underlying type Any)
required: T
b2.f(a2)
^
Why? (Besides that, the approach has the disadvantage of incurring runtime costs (I think) because of the creation and destruction of a pair.)
b match {
case b : C[t] => b.f(a.asInstanceOf[t])
}
This works. But enclosing the code with a match makes the code much less readable. (And it also is too complicated for Intelli/J.)
The cleanest solution is, IMO, the one you found with the type-capture pattern match. You can make it concise, and hopefully readable, by integrating the pattern directly inside your for comprehension, as follows:
for ((a, b: C[t]) <- list) {
b.f(a.asInstanceOf[t])
}
Fiddle: http://www.scala-js-fiddle.com/gist/b9030033133ee94e8c18ad772f3461a0
If you are not in a for comprehension already, unfortunately the corresponding pattern assignment does not work:
val (c, d: C[t]) = (a, b)
d.f(c.asInstanceOf[t])
That's because t is not in scope anymore on the second line. In that case, you would have to use the full pattern matching.
Maybe I'm confused about what you are trying to achieve, but this compiles:
case class C[T](x:T) {
def f(t:T) = println(t)
type ValueType = T
}
type CP[T] = (T, C[T])
val list = List[CP[T forSome {type T}]](1 -> C(2), "hello" -> C("goodbye"))
for ((a,b) <- list) {
b.f(a)
}
Edit
If the type of the list itself is out of your control, you can still cast it to this "correct" type.
case class C[T](x:T) {
def f(t:T) = println(t)
type ValueType = T
}
val list = List(1 -> C(2), "hello" -> C("goodbye"))
type CP[T] = (T, C[T])
for ((a,b) <- list.asInstanceOf[List[CP[T forSome { type T }]]]) {
b.f(a)
}
Great question! Lots to learn here about Scala.
Other answers and comments have already addressed most of the issues here, but I'd like to address a few additional points.
You asked why this variant doesn't work:
val (a2,b2) = (a,b).asInstanceOf[(T,C[T]) forSome {type T}]
b2.f(a2)
You aren't the only person who's been surprised by this; see e.g. this recent very similar issue report: SI-9899.
As I wrote there:
I think this is working as designed as per SLS 6.1: "The following skolemization rule is applied universally for every expression: If the type of an expression would be an existential type T, then the type of the expression is assumed instead to be a skolemization of T."
Basically, every time you write a value-level expression that the compiler determines to have an existential type, the existential type is instantiated. b2.f(a2) has two subexpressions with existential type, namely b2 and a2, so the existential gets two different instantiations.
As for why the pattern-matching variant works, there isn't explicit language in SLS 8 (Pattern Matching) covering the behavior of existential types, but 6.1 doesn't apply because a pattern isn't technically an expression, it's a pattern. The pattern is analyzed as a whole and any existential types inside only get instantiated (skolemized) once.
As a postscript, note that yes, when you play in this area, the error messages you get are often confusing or misleading and ought to be improved. See for example https://github.com/scala/scala-dev/issues/205
A wild guess, but is it possible that you need something like this:
case class C[+T](x:T) {
def f[A >: T](t: A) = println(t)
}
val list = List(1 -> C(2), "hello" -> C("goodbye"))
for ((a,b) <- list) {
b.f(a)
}
?
It will type check.
I'm not quite sure what "runtime guarantee" means here, usually it means that you are trying to fool type system (e.g. with asInstanceOf), but then all bets are off and you shouldn't expect type system to be of any help.
UPDATE
Just for the illustration why type casting is an evil:
case class C[T <: Int](x:T) {
def f(t: T) = println(t + 1)
}
val list = List("hello" -> C(2), 2 -> C(3))
for ((a, b: C[t]) <- list) {
b.f(a.asInstanceOf[t])
}
It compiles and fails at runtime (not surprisingly).
UPDATE2
Here's what generated code looks like for the last snippet (with C[t]):
...
val a: Object = x1._1();
val b: Test$C = x1._2().$asInstanceOf[Test$C]();
if (b.ne(null))
{
<synthetic> val x2: Test$C = b;
matchEnd4({
x2.f(scala.Int.unbox(a));
scala.runtime.BoxedUnit.UNIT
})
}
...
Type t simply vanished (as it should have been) and Scala is trying to convert a to an upper bound of T in C, i.e. Int. If there is no upper bound it's going to be Any (but then method f is nearly useless unless you cast again or use something like println which takes Any).

Compile-time Reflection of Nested List: typecheck List[List[Int]] returns List[List[...]]?

I'm using macro annotations to inspect the fields of a class and add a member based on those fields.
e.g.
#AddVal
class A(x: Int)
expands to
class A(x: Int){
val get: Int = x
}
After extracting theValDef, it's tpe field still null so to get the type I have two options:
1) If I call .toString on the type tree, I can see the type, but now I've lost some type-safety
2) If I use c.typecheck on the type tree, I can get the type, but only if it's 1 level deep. List[List[Int]] comes back as List[List[...]]
val fieldType = c.typecheck(q"type T = ${f.tpt}") match {
case x # TypeDef(mods, name, tparams, rhs) => rhs.tpe
}
So, is there a way to recursively typecheck polytypes?
I tried typechecking rhs again but I got The argument types of an anonymous function must be fully known and I'm not sure how to resolve that.
Thanks for taking a look,
Julian
I incorrectly attributed this error to the macro, when in fact there was another underlying macro (a type provider macro) that was failing to provide the proper nested type (in this case the Int).

Demystifying a function definition

I am new to Scala, and I hope this question is not too basic. I couldn't find the answer to this question on the web (which might be because I don't know the relevant keywords).
I am trying to understand the following definition:
def functionName[T <: AnyRef](name: Symbol)(range: String*)(f: T => String)(implicit tag: ClassTag[T]): DiscreteAttribute[T] = {
val r = ....
new anotherFunctionName[T](name.toString, f, Some(r))
}
First , why is it defined as def functionName[...](...)(...)(...)(...)? Can't we define it as def functionName[...](..., ..., ..., ...)?
Second, how does range: String* from range: String?
Third, would it be a problem if implicit tag: ClassTag[T] did not exist?
First , why is it defined as def functionName...(...)(...)(...)? Can't we define it as def functionName[...](..., ..., ..., ...)?
One good reason to use currying is to support type inference. Consider these two functions:
def pred1[A](x: A, f: A => Boolean): Boolean = f(x)
def pred2[A](x: A)(f: A => Boolean): Boolean = f(x)
Since type information flows from left to right if you try to call pred1 like this:
pred1(1, x => x > 0)
type of the x => x > 0 cannot be determined yet and you'll get an error:
<console>:22: error: missing parameter type
pred1(1, x => x > 0)
^
To make it work you have to specify argument type of the anonymous function:
pred1(1, (x: Int) => x > 0)
pred2 from the other hand can be used without specifying argument type:
pred2(1)(x => x > 0)
or simply:
pred2(1)(_ > 0)
Second, how does range: String* from range: String?
It is a syntax for defining Repeated Parameters a.k.a varargs. Ignoring other differences it can be used only on the last position and is available as a scala.Seq (here scala.Seq[String]). Typical usage is apply method of the collections types which allows for syntax like SomeDummyCollection(1, 2, 3). For more see:
What does `:_*` (colon underscore star) do in Scala?
Scala variadic functions and Seq
Is there a difference in Scala between Seq[T] and T*?
Third, would it be a problem if implicit tag: ClassTag[T] did not exist?
As already stated by Aivean it shouldn't be the case here. ClassTags are automatically generated by the compiler and should be accessible as long as the class exists. In general case if implicit argument cannot be accessed you'll get an error:
scala> import scala.concurrent._
import scala.concurrent._
scala> val answer: Future[Int] = Future(42)
<console>:13: error: Cannot find an implicit ExecutionContext. You might pass
an (implicit ec: ExecutionContext) parameter to your method
or import scala.concurrent.ExecutionContext.Implicits.global.
val answer: Future[Int] = Future(42)
Multiple argument lists: this is called "currying", and enables you to call a function with only some of the arguments, yielding a function that takes the rest of the arguments and produces the result type (partial function application). Here is a link to Scala documentation that gives an example of using this. Further, any implicit arguments to a function must be specified together in one argument list, coming after any other argument lists. While defining functions this way is not necessary (apart from any implicit arguments), this style of function definition can sometimes make it clearer how the function is expected to be used, and/or make the syntax for partial application look more natural (f(x) rather than f(x, _)).
Arguments with an asterisk: "varargs". This syntax denotes that rather than a single argument being expected, a variable number of arguments can be passed in, which will be handled as (in this case) a Seq[String]. It is the equivalent of specifying (String... range) in Java.
the implicit ClassTag: this is often needed to ensure proper typing of the function result, where the type (T here) cannot be determined at compile time. Since Scala runs on the JVM, which does not retain type information beyond compile time, this is a work-around used in Scala to ensure information about the type(s) involved is still available at runtime.
Check currying:Methods may define multiple parameter lists. When a method is called with a fewer number of parameter lists, then this will yield a function taking the missing parameter lists as its arguments.
range:String* is the syntax for varargs
implicit TypeTag parameter in Scala is the alternative for Class<T> clazzparameter in Java. It will be always available if your class is defined in scope. Read more about type tags.

What is the point of multiple parameter clauses in function definitions in Scala?

I'm trying to understand the point of this language feature of multiple parameter clauses and why you would use it.
Eg, what's the difference between these two functions really?
class WTF {
def TwoParamClauses(x : Int)(y: Int) = x + y
def OneParamClause(x: Int, y : Int) = x + y
}
>> val underTest = new WTF
>> underTest.TwoParamClauses(1)(1) // result is '2'
>> underTest.OneParamClause(1,1) // result is '2'
There's something on this in the Scala specification at point 4.6. See if that makes any sense to you.
NB: the spec calls these 'parameter clauses', but I think some people may also call them 'parameter lists'.
Here are three practical uses of multiple parameter lists,
To aid type inference. This is especially useful when using higher order methods. Below, the type parameter A of g2 is inferred from the first parameter x, so the function arguments in the second parameter f can be elided,
def g1[A](x: A, f: A => A) = f(x)
g1(2, x => x) // error: missing parameter type for argument x
def g2[A](x: A)(f: A => A) = f(x)
g2(2) {x => x} // type is inferred; also, a nice syntax
For implicit parameters. Only the last parameter list can be marked implicit, and a single parameter list cannot mix implicit and non-implicit parameters. The definition of g3 below requires two parameter lists,
// analogous to a context bound: g3[A : Ordering](x: A)
def g3[A](x: A)(implicit ev: Ordering[A]) {}
To set default values based on previous parameters,
def g4(x: Int, y: Int = 2*x) {} // error: not found value x
def g5(x: Int)(y: Int = 2*x) {} // OK
TwoParamClause involves two method invocations while the OneParamClause invokes the function method only once. I think the term you are looking for is currying. Among the many use cases, it helps you to breakdown the computation into small steps. This answer may convince you of usefulness of currying.
There is a difference between both versions concerning type inference. Consider
def f[A](a:A, aa:A) = null
f("x",1)
//Null = null
Here, the type A is bound to Any, which is a super type of String and Int. But:
def g[A](a:A)(aa:A) = null
g("x")(1)
error: type mismatch;
found : Int(1)
required: java.lang.String
g("x")(1)
^
As you see, the type checker only considers the first argument list, so A gets bound to String, so the Int value for aa in the second argument list is a type error.
Multiple parameter lists can help scala type inference for more details see: Making the most of Scala's (extremely limited) type inference
Type information does not flow from left to right within an argument list, only from left to right across argument lists. So, even though Scala knows the types of the first two arguments ... that information does not flow to our anonymous function.
...
Now that our binary function is in a separate argument list, any type information from the previous argument lists is used to fill in the types for our function ... therefore we don't need to annotate our lambda's parameters.
There are some cases where this distinction matters:
Multiple parameter lists allow you to have things like TwoParamClauses(2); which is automatically-generated function of type Int => Int which adds 2 to its argument. Of course you can define the same thing yourself using OneParamClause as well, but it will take more keystrokes
If you have a function with implicit parameters which also has explicit parameters, implicit parameters must be all in their own parameter clause (this may seem as an arbitrary restriction but is actually quite sensible)
Other than that, I think, the difference is stylistic.