Add a check for correct Array shape in quotes.reflect.ClassOfConstant

[jchyb]

Closes #21916
I tried to supply the ClassOfConstant with multiple other broken Types, but I was unable to break it beyond the linked issue, so I ended up adding the check for only that one case. This makes sense - the backend (and thus erasure) needs to know if the Array type parameter is a primitive type, but in other cases the erasure phase needs to know only the class, without the type parameters.

It's impossible to call classOf through the quoted code ('{classOf[t]} with a boundless t will error out), so we don't need that additional check there.

There does appear to be an issue with being able to set '{List[Array]} resulting in a crash, but that is beyond the scope of this fix - I will prepare a separate issue for that (edit: reported here).

[hamzaremmal]

Why a pattern matching if both branches have a similar result ?

[jchyb]

I forgot about case AppliedType(_, _) where we would not check isArray. I've adjusted this to be clearer anyway.

[hamzaremmal]

We should follow the specification of classOf here. Even if it doesn't create any issues in later phases, we should not allow types that are not proper types (*-kind).

[hamzaremmal]

@jchyb The current check for Predef.classOf:

scala3/compiler/src/dotty/tools/dotc/typer/Checking.scala

Line 1129 in d4421d0

def checkClassType(tp: Type, pos: SrcPos, traitReq: Boolean, stablePrefixReq: Boolean, refinementOK: Boolean = false)(using Context): Type =

[jchyb]

I agree, and in a ideal world that would be the case, but that's something I am personally very conflicted about - adding a check like this would completely break Izumi reflect (they always strip the type arguments, for whatever reason), and thus any zio libraries that use it (I unfortunately did also end up using Izumi-reflect in other migrations, since it's pretty much the only alternative to the deprecated TypeTag from scala 2). I wanted the change here to have as small of a footprint as possible, while still being helpful.

[hamzaremmal]

@jchyb Can you please open an issue to track the fact that we need to add the "correct" check eventually and add it to the backlog.