Should infix work on enum case?

[KacperFKorban]

Compiler version

7ba88ad

Minimized code

enum Expr1:
  infix case Add1[L, R](l: L, r: R)

Output

-- Error: /home/kpi/wiadro/infix_things.scala:2:8 ------------------------------
2 |  infix case Add1[L, R](l: L, r: R)
  |        ^^^^
  |        end of statement expected but 'case' found
-- [E145] Syntax Error: /home/kpi/wiadro/infix_things.scala:1:5 ----------------
1 |enum Expr1:
  |     ^^^^^
  |     Enumerations must contain at least one case
  |
  | longer explanation available when compiling with `-explain`
-- [E006] Not Found Error: /home/kpi/wiadro/infix_things.scala:2:2 -------------
2 |  infix case Add1[L, R](l: L, r: R)
  |  ^^^^^
  |  Not found: infix
  |
  | longer explanation available when compiling with `-explain`

Expectation

infix should be allowed on enum cases (that declare class types)

Other info

Other similar issues:

• relevant discussion about infix enum: Sould infix work on enum? #18933
• discussion about supported infix constructs: infix keyword is allowed where it has no effect, and has undocumented uses #17738

[hamzaremmal]

spree ?

[KacperFKorban]

At first sight, it seems just right, but there might be some edge cases I don't see now. CC @mbovel
There is also a question of whether it should even be supported. I felt like the decision in #18933 was that it should. Is there a spec for infix things?

[KacperFKorban]

@sjrd Do you know if we want to support this?

[sjrd]

In an ideal world, List would be an enum with a case Nil and a case ::(head, tail). In that world, we would definitely want :: to be infix. So I think we want to support infix case definitions for class cases.

[mbovel]

This issue was picked for the Scala Issue Spree of tomorrow, Monday, December 2nd. @KacperFKorban, @kasiaMarek and @hamzaremmal will be working on it.

[kasiaMarek]

As described here allowing infix (a soft modifier) before case changes the specification.