K2: Fix processing inference lower bound NullableType <: T & Any

Previously, it was led to plainly adding NullableType <: T constraint
which silently led to successful call completion.
What is suggested is just marking such initial constraint
as unsuccessful.

In K1, the error was reported just via additional type checking
mechanism being run after call completion.

^KT-58665 Fixed

compiler/testData/diagnostics/tests/explicitDefinitelyNotNullableViaIntersection/approximation.fir.kt

@@ -1,14 +0,0 @@
-// !LANGUAGE: +DefinitelyNonNullableTypes
-
-fun <T> foo(x: T, y: T & Any) = x!!
-
-fun main() {
-    foo<String>("", "").length
-    foo<String>("", <!NULL_FOR_NONNULL_TYPE!>null<!>).length
-    foo<String?>(null, "").length
-    foo<String?>(null, null).length
-
-    foo("", "").length
-    foo("", null).length
-    foo(null, "").length
-}

compiler/testData/diagnostics/tests/explicitDefinitelyNotNullableViaIntersection/approximation.kt

@@ -1,3 +1,4 @@
+// FIR_IDENTICAL
 // !LANGUAGE: +DefinitelyNonNullableTypes
 
 fun <T> foo(x: T, y: T & Any) = x!!

compiler/testData/diagnostics/tests/explicitDefinitelyNotNullableViaIntersection/inference.fir.kt

@@ -28,6 +28,6 @@ fun <F> main(x: F, y: F, z: F, w: F, m: F) {
     w1.foo()
     w2.foo()
 
-    expectNN(m)
+    expectNN(<!ARGUMENT_TYPE_MISMATCH!>m<!>)
     expectNN(m!!)
 }

compiler/testData/diagnostics/tests/explicitDefinitelyNotNullableViaIntersection/simple.fir.kt

@@ -1,14 +0,0 @@
-// !LANGUAGE: +DefinitelyNonNullableTypes
-
-fun <T> foo(x: T, y: T & Any): T & Any = x ?: y
-
-fun main() {
-    foo<String>("", "").length
-    foo<String>("", <!NULL_FOR_NONNULL_TYPE!>null<!>).length
-    foo<String?>(null, "").length
-    foo<String?>(null, null).length
-
-    foo("", "").length
-    foo("", null).length
-    foo(null, "").length
-}

compiler/testData/diagnostics/tests/explicitDefinitelyNotNullableViaIntersection/simple.kt

@@ -1,3 +1,4 @@
+// FIR_IDENTICAL
 // !LANGUAGE: +DefinitelyNonNullableTypes
 
 fun <T> foo(x: T, y: T & Any): T & Any = x ?: y