bpo-26701: Add documentation for __trunc__ (GH-6022)

`int` fails back to `__trunc__` is `__int__` isn't defined, so cover
that in the docs.

Doc/library/functions.rst

@@ -731,8 +731,11 @@ are always available.  They are listed here in alphabetical order.
 
    Return an integer object constructed from a number or string *x*, or return
    ``0`` if no arguments are given.  If *x* is a number, return
-   :meth:`x.__int__() <object.__int__>`.  For floating point numbers, this
-   truncates towards zero.
+   :meth:`x.__int__() <object.__int__>`. If *x* defines
+   :meth:`x.__trunc__() <object.__trunc__>` but not
+   :meth:`x.__int__() <object.__int__>`, then return
+   if :meth:`x.__trunc__() <object.__trunc__>`.  For floating point numbers,
+   this truncates towards zero.
 
    If *x* is not a number or if *base* is given, then *x* must be a string,
    :class:`bytes`, or :class:`bytearray` instance representing an :ref:`integer

Doc/library/math.rst

@@ -203,7 +203,7 @@ Number-theoretic and representation functions
 
    Return the :class:`~numbers.Real` value *x* truncated to an
    :class:`~numbers.Integral` (usually an integer). Delegates to
-   ``x.__trunc__()``.
+   :meth:`x.__trunc__() <object.__trunc__>`.
 
 
 Note that :func:`frexp` and :func:`modf` have a different call/return pattern

Doc/reference/datamodel.rst

@@ -2377,6 +2377,15 @@ left undefined.
    of the appropriate type.
 
 
+.. method:: object.__trunc__(self)
+
+   Called to implement :meth:`math.trunc`.  Should return the value of the
+   object truncated to a :class:`numbers.Integral` (typically an
+   :class:`int`).  If a class defines :meth:`__trunc__` but not
+   :meth:`__int__`, then :meth:`__trunc__` is called to implement the
+   built-in function :func:`int`.
+
+
 .. method:: object.__index__(self)
 
    Called to implement :func:`operator.index`, and whenever Python needs to