wdyt about only warning under -Xsource:3?

tired: warn under -Xsource:3

wired: default to -Xsource:3-cross -Werror

The question whether it's worth doing anything on Scala 2 where there is no development branch (nothing can change) made me second-guess whether this is worth doing on Scala 3, so I'll comment on the discussion. I see discussions don't get a link when mentioned.

Coincidentally scala/scala3#21681 named tuples forces the issue. That got a mention on the discussion (the discussion between myself).

Adds -Xsource:3 for the deprecation; demotes -Wmultiarg-infix because it does not warn in Dotty and also does not produce a named tuple, so that it is unlikely to require any messaging ever, but retained as an opt-in warning for anyone who feels strongly about it and doesn't use a DSL that requires it.

Requesting Seth as a reviewer and cashing in a favor for (e: Unit) syntax.

I have to review the last commit, or ask Seth. I just noticed:

importInfos(ctx.unit) ::= (info, keyword, ctx.owner) : @nowarn

where it's taken as a tuple, as intended. So this commit makes the warning opt-in and the nowarn will be unnecessary.

OK, I see 3.5.2 prefers multiarg infix over tupled in the overloaded case. 2.13.15 does the same.

In other words, X op (1, 2, 3) is not syntactically a tuple; the operand is tupled as an adaptation, if necessary.

"How do I warn about that, again?"

➜  ~ scala -Xlint:adapted-args
Welcome to Scala 2.13.15 (OpenJDK 64-Bit Server VM, Java 23).
Type in expressions for evaluation. Or try :help.

scala> def f[A](x: A) = x
def f[A](x: A): A

scala> f(1,2,3)
        ^
       warning: adapted the argument list to the expected 3-tuple: add additional parens instead
               signature: f[A](x: A): A
         given arguments: 1, 2, 3
        after adaptation: f((1, 2, 3): (Int, Int, Int)) [quickfixable]
val res0: (Int, Int, Int) = (1,2,3)

scala> object X { def f[A](x: A) = x.getClass }
object X

scala> X f(1,2,3)
val res1: Class[_ <: (Int, Int, Int)] = class scala.Tuple3

scala>

Aha, it does not warn for the infix syntax!

As a footnote, I remember saying that multiarg infix specifically means multiarg, which is why nilary and unary work correctly; for example, () == (). But by the same token (pun), () op () should be ().op() first and only adapted to ().op(()), and similarly for x op (y). x op (42) is never adapted where Tuple1[A] is expected, which seems wrong.

I can't delete the @nowarn for multiarg infix on my other PR until this is in; or, I'll just cherry-pick it.