fix regression with uint constant losing abstract type (#24105)

fixes #24104, refs #23955

The line `result.typ = dstTyp` added in #23955 changes the type of
`result`, which was the type of `n` due to the argument passed to
`newIntNodeT`, to the abstract type skipped `dstTyp`. The line is
removed to just keep the type as abstract.

(cherry picked from commit 6d362e0ffe)

compiler/semfold.nim

@@ -412,7 +412,6 @@ proc foldConv(n, a: PNode; idgen: IdGenerator; g: ModuleGraph; check = false): P
       if dstTyp.kind in {tyUInt..tyUInt64}:
         result = newIntNodeT(maskBytes(val, getSize(g.config, dstTyp)), n, idgen, g)
         result.transitionIntKind(nkUIntLit)
-        result.typ = dstTyp
       else:
         if check: rangeCheck(n, val, g)
         result = newIntNodeT(val, n, idgen, g)

tests/int/t1.nim

@@ -42,3 +42,10 @@ block: # bug #23954
   doAssert testRT_u8 == 7
   const testCT_u8 : uint8 = 0x107.uint8
   doAssert testCT_u8 == 7
+
+block: # issue #24104
+  type P = distinct uint  # uint, uint8, uint16, uint32, uint64 
+  let v = 0.P
+  case v
+  of 0.P: discard
+  else:   discard