Adding Id.apply

[BalmungSan]

I hope there isn't a reason for not doing this.

[joroKr21]

You could just do a: Id[A]

[BalmungSan]

@joroKr21 sure, but a lot of times I have wanted to use Id to then use map as a pipe of pure functions to pure values. But then if I have to do (foo: Id[Foo]).map(f) looks worse and usually breaks this chain idea I had in mind in the first place.
The idea of this PR is to allow us to do Id(foo).map(f)

[satorg]

@BalmungSan another way with help of the existing Cats syntax could be a.pure[Id]:

import cats._
import cats.syntax.all._

val a = 123
println(a.pure[Id].map(_ + 456))

https://scastie.scala-lang.org/satorg/Qq3pz0kaT8CLJqfTye2VBA/17

[joroKr21]

Since Scala 2.13 there is also a standard library pipe syntax in scala.util.chaining btw

[BalmungSan]

@satorg oh nice, I kind of always forget about it.
@joroKr21 right, I personally don't like the names the used but yeah that is a valid point.

I guess the point of the PR is not really if this is the best way or not to access it. But, rather why not having the Id object?
Even more, I think this could serve as the initial step of a bigger change of moving Id into cats.data and moving the typeclass instances of Id into its companion object. The problem is that such changes are source and binary breaking so they must wait and I am not sure if everyone agree if those make sense.

[joroKr21]

Yeah it's definitely a valid point to discuss on it's own merits whether Id.apply should exist or not.
I'm just giving you some alternatives you can apply today.

Regarding the type class instances, that wouldn't work in Scala 2 anyway because type aliases can't have a companion object. I.e. you can have an object Id, it's just not a companion. A completely separate question is if Id should be an opaque type in a hypothetical Scala 3 only version of Cats.

[tpolecat]

I like this change. It's trivial and it provides an ergonomic improvement that I would use from time to time.

[satorg]

I personally don't mind this change. Just to note: there's IdSuite as well. I wonder would it make sense to put a small test for this change there? Something like

property("Id#apply") {
  forAll { (a: Int) =>
    assert(Id(a) == (a: Id[Int])) // or `== a.pure[Id]` or even just `== a`
  }
}

[BalmungSan]

Uhm that error is not related to my change, is it?

[tpolecat]

I restarted the build.

[rossabaker]

There are other ways, but I've looked for this way before. It's intuitive to me.

I wish it lived in cats.data, but it's not easy to get there from here in Cats 2.