typeToString can now show (recursively) resolved type aliases; fixes #8569 #8083 #8570 (#11678)

* nested typeToString
* typeToString: preferResolved
* add test
* fix test
* preferMixed
* fix tests

tests/errmsgs/tsigmatch.nim

@@ -36,7 +36,7 @@ tsigmatch.nim(143, 13) Error: type mismatch: got <array[0..0, proc (x: int){.gcs
 but expected one of:
 proc takesFuncs(fs: openArray[proc (x: int) {.gcsafe, locks: 0.}])
   first type mismatch at position: 1
-  required type for fs: openarray[proc (x: int){.closure, gcsafe, locks: 0.}]
+  required type for fs: openArray[proc (x: int){.closure, gcsafe, locks: 0.}]
   but expression '[proc (x: int) {.gcsafe, locks: 0.} = echo [x]]' is of type: array[0..0, proc (x: int){.gcsafe, locks: 0.}]
 
 expression: takesFuncs([proc (x: int) {.gcsafe, locks: 0.} = echo [x]])

tests/errmsgs/twrong_at_operator.nim

@@ -6,7 +6,7 @@ twrong_at_operator.nim(22, 30) Error: type mismatch: got <array[0..0, type int]>
 but expected one of:
 proc `@`[T](a: openArray[T]): seq[T]
   first type mismatch at position: 1
-  required type for a: openarray[T]
+  required type for a: openArray[T]
   but expression '[int]' is of type: array[0..0, type int]
 proc `@`[IDX, T](a: sink array[IDX, T]): seq[T]
   first type mismatch at position: 1

tests/metatype/ttypetraits2.nim

@@ -14,3 +14,32 @@ block: # isNamedTuple
   doAssert not Foo3.isNamedTuple
   doAssert not Foo4.isNamedTuple
   doAssert not (1,).type.isNamedTuple
+
+proc typeToString*(t: typedesc, prefer = "preferTypeName"): string {.magic: "TypeTrait".}
+  ## Returns the name of the given type, with more flexibility than `name`,
+  ## and avoiding the potential clash with a variable named `name`.
+  ## prefer = "preferResolved" will resolve type aliases recursively.
+  # Move to typetraits.nim once api stabilized.
+
+block: # typeToString
+  type MyInt = int
+  type
+    C[T0, T1] = object
+  type C2=C # alias => will resolve as C
+  type C2b=C # alias => will resolve as C (recursively)
+  type C3[U,V] = C[V,U]
+  type C4[X] = C[X,X]
+  template name2(T): string = typeToString(T, "preferResolved")
+  doAssert MyInt.name2 == "int"
+  doAssert C3[MyInt, C2b].name2 == "C3[int, C]"
+    # C3 doesn't get resolved to C, not an alias (nor does C4)
+  doAssert C2b[MyInt, C4[cstring]].name2 == "C[int, C4[cstring]]"
+  doAssert C4[MyInt].name2 == "C4[int]"
+  when BiggestFloat is float and cint is int:
+    doAssert C2b[cint, BiggestFloat].name2 == "C3[int, C3[float, int32]]"
+
+  template name3(T): string = typeToString(T, "preferMixed")
+  doAssert MyInt.name3 == "MyInt{int}"
+  doAssert (tuple[a: MyInt, b: float]).name3 == "tuple[a: MyInt{int}, b: float]"
+  doAssert (tuple[a: C2b[MyInt, C4[cstring]], b: cint, c: float]).name3 ==
+    "tuple[a: C2b{C}[MyInt{int}, C4[cstring]], b: cint{int32}, c: float]"

tests/openarray/t8259.nim

@@ -1,5 +1,5 @@
 discard """
-  errormsg: "invalid type: 'openarray[int]' for result"
+  errormsg: "invalid type: 'openArray[int]' for result"
   line: 6
 """