case of and intersection types

[JaroslavTulach]

Pull Request Description

Fixes #11828 by enhancing the case of nodes to recognize additional types of EnsoMultiValue.

Important Notes

Checklist

Please ensure that the following checklist has been satisfied before submitting the PR:

• All code follows the
  Scala,
  Java,
• Unit tests have been written where possible.

[JaroslavTulach]

This shows a small nuance. Previously there was no difference between that and n. Now there is as 571cce6 documents.

[4e6]

Please document all non-trivial logic

[Akirathan]

Looks good altogether. I have two questions though. The first one about incosistency is more important:

(1)

So type pattern branches in case of expressions don't do automatic conversions, as opposed to standard method calls.
i.e.,

a = A:A
case a of
    b: B -> "Is a B"

will not match. In other words, case expressions uncover the hidden parts of multivalue, but they don't uncover the types that the scrutinee can be cast to. I find this slightly confusing. Consider this code:

method (b:B) = b.b

main =
    a = A:A
    ax = A:A&X
    res1 = case ax of
        b: B -> b.b
        _ -> "Else"
    res2 = method ax
    [res1, res2]

res1 will be "Else", while res2 will be "b". The same thing will happen if ax is replaced with a in the case expression and method call. What is the reason for this incosistency?

(2)

Any plans to support intersection types in type patterns as well?

case ax of
    _: A&X -> ...

fails with:

/home/pavel/tmp/tmp.enso:27:12: error: Syntax is not supported yet: qualifiedNameSegment.
   27 |         _: A&X -> "Is A&X"
      |            ^~~

[JaroslavTulach]

So type pattern branches in case of expressions don't do automatic conversions

I've asked and @radeusgd says that the pattern branches should be consistent with is_a function. E.g. no conversions.

Any plans to support intersection types in type patterns as well?

Yes, reported as #11864