Revert "make default values typed in proc AST same as param sym AST" (#…

…24191)

Reverts #24184, reopens #12942, reopens #19118

#24184 seems to have caused a regression in
https://github.com/c-blake/thes and
https://github.com/c-blake/bu/blob/main/rp.nim#L84 reproducible with
`git clone https://github.com/c-blake/cligen; git clone
https://github.com/c-blake/thes; cd thes; nim c -p=../cligen thes`.
Changing the `const` to `let` makes it compile.

A minimization that is probably the same issue is:

```nim
const a: seq[string] = @[]

proc foo(x = a) =
  echo typeof(x)
  echo x

import macros

macro resemFoo() =
  result = getImpl(bindSym"foo")

block:
  resemFoo() # Error: cannot infer the type of parameter 'x'
```

This should be a regression test in a future reimplementation of #24184.

compiler/semtypes.nim

@@ -1409,8 +1409,6 @@ proc semProcTypeNode(c: PContext, n, genericParams: PNode,
         elif typ.kind == tyStatic:
           def = semConstExpr(c, def)
           def = fitNode(c, typ, def, def.info)
-      # keep proc AST updated
-      a[^1] = def
 
     if not hasType and not hasDefault:
       if isType: localError(c.config, a.info, "':' expected")

tests/errmsgs/tunknown_named_parameter.nim

@@ -10,8 +10,8 @@ func rsplit(s: string; sep: string; maxsplit: int = -1): seq[string]
   first type mismatch at position: 2
   required type for sep: string
   but expression '{':'}' is of type: set[char]
-func rsplit(s: string; seps: set[char] = {' ', '\t', '\v', '\r', '\n', '\f'};
-            maxsplit: int = -1): seq[string]
+func rsplit(s: string; seps: set[char] = Whitespace; maxsplit: int = -1): seq[
+    string]
   first type mismatch at position: 3
   unknown named parameter: maxsplits