documented overload disambiguation [backport] (#18865)

* documented overload disambiguation [backport]

* Update doc/manual.rst

Co-authored-by: konsumlamm <44230978+konsumlamm@users.noreply.github.com>

* documented overload disambiguation [backport]

* documented overload disambiguation [backport]

Co-authored-by: konsumlamm <44230978+konsumlamm@users.noreply.github.com>

doc/manual.rst

@@ -2708,6 +2708,24 @@ a parameter typed as `untyped` (for unresolved expressions) or the type class
   assert toSeq2(items(@[1,2])) == @[1, 2] # but items(@[1,2]) is
 
 
+Overload disambiguation
+=======================
+
+For routine calls "overload resolution" is performed. There is a weaker form of
+overload resolution called *overload disambiguation* that is performed when an
+overloaded symbol is used in a context where there is additional type information
+available. Let `p` be an overloaded symbol. These contexts are:
+
+- In a function call `q(..., p, ...)` when the corresponding formal parameter
+  of `q` is a `proc` type. If `q` itself is overloaded then the cartesian product
+  of every interpretation of `q` and `p` must be considered.
+- In an object constructor `Obj(..., field: p, ...)` when `field` is a `proc`
+  type. Analogous rules exist for array/set/tuple constructors.
+- In a declaration like `x: T = p` when `T` is a `proc` type.
+
+As usual, ambiguous matches produce a compile-time error.
+
+
 Statements and expressions
 ==========================