bpo-17799: Explain real behaviour of sys.settrace and sys.setprofile (#4056)

diff --git a/Doc/c-api/init.rst b/Doc/c-api/init.rst
index 4ee61e8..86faf47 100644
--- a/Doc/c-api/init.rst
+++ b/Doc/c-api/init.rst
@@ -1300,7 +1300,6 @@
    | :const:`PyTrace_C_RETURN`    | Function object being called.        |
    +------------------------------+--------------------------------------+
 
-
 .. c:var:: int PyTrace_CALL
 
    The value of the *what* parameter to a :c:type:`Py_tracefunc` function when a new
@@ -1357,16 +1356,18 @@
    function as its first parameter, and may be any Python object, or *NULL*.  If
    the profile function needs to maintain state, using a different value for *obj*
    for each thread provides a convenient and thread-safe place to store it.  The
-   profile function is called for all monitored events except the line-number
-   events.
+   profile function is called for all monitored events except :const:`PyTrace_LINE`
+   and :const:`PyTrace_EXCEPTION`.
 
 
 .. c:function:: void PyEval_SetTrace(Py_tracefunc func, PyObject *obj)
 
    Set the tracing function to *func*.  This is similar to
    :c:func:`PyEval_SetProfile`, except the tracing function does receive line-number
-   events.
-
+   events and does not receive any event related to C function objects being called. Any
+   trace function registered using :c:func:`PyEval_SetTrace` will not receive
+   :const:`PyTrace_C_CALL`, :const:`PyTrace_C_EXCEPTION` or :const:`PyTrace_C_RETURN`
+   as a value for the *what* parameter.
 
 .. _advanced-debugging:
 
diff --git a/Doc/library/sys.rst b/Doc/library/sys.rst
index 54281a3..ab08f61 100644
--- a/Doc/library/sys.rst
+++ b/Doc/library/sys.rst
@@ -1084,13 +1084,38 @@
    Set the system's profile function, which allows you to implement a Python source
    code profiler in Python.  See chapter :ref:`profile` for more information on the
    Python profiler.  The system's profile function is called similarly to the
-   system's trace function (see :func:`settrace`), but it isn't called for each
-   executed line of code (only on call and return, but the return event is reported
-   even when an exception has been set).  The function is thread-specific, but
-   there is no way for the profiler to know about context switches between threads,
-   so it does not make sense to use this in the presence of multiple threads. Also,
+   system's trace function (see :func:`settrace`), but it is called with different events,
+   for example it isn't called for each executed line of code (only on call and return,
+   but the return event is reported even when an exception has been set). The function is
+   thread-specific, but there is no way for the profiler to know about context switches between
+   threads, so it does not make sense to use this in the presence of multiple threads. Also,
    its return value is not used, so it can simply return ``None``.
 
+   Profile functions should have three arguments: *frame*, *event*, and
+   *arg*. *frame* is the current stack frame.  *event* is a string: ``'call'``,
+   ``'return'``, ``'c_call'``, ``'c_return'``, or ``'c_exception'``. *arg* depends
+   on the event type.
+
+   The events have the following meaning:
+
+   ``'call'``
+      A function is called (or some other code block entered).  The
+      profile function is called; *arg* is ``None``.
+
+   ``'return'``
+      A function (or other code block) is about to return.  The profile
+      function is called; *arg* is the value that will be returned, or ``None``
+      if the event is caused by an exception being raised.
+
+   ``'c_call'``
+      A C function is about to be called.  This may be an extension function or
+      a built-in.  *arg* is the C function object.
+
+   ``'c_return'``
+      A C function has returned. *arg* is the C function object.
+
+   ``'c_exception'``
+      A C function has raised an exception.  *arg* is the C function object.
 
 .. function:: setrecursionlimit(limit)
 
@@ -1137,8 +1162,8 @@
 
    Trace functions should have three arguments: *frame*, *event*, and
    *arg*. *frame* is the current stack frame.  *event* is a string: ``'call'``,
-   ``'line'``, ``'return'``, ``'exception'``, ``'c_call'``, ``'c_return'``, or
-   ``'c_exception'``, ``'opcode'``. *arg* depends on the event type.
+   ``'line'``, ``'return'``, ``'exception'`` or ``'opcode'``.  *arg* depends on
+   the event type.
 
    The trace function is invoked (with *event* set to ``'call'``) whenever a new
    local scope is entered; it should return a reference to a local trace
@@ -1175,16 +1200,6 @@
       tuple ``(exception, value, traceback)``; the return value specifies the
       new local trace function.
 
-   ``'c_call'``
-      A C function is about to be called.  This may be an extension function or
-      a built-in.  *arg* is the C function object.
-
-   ``'c_return'``
-      A C function has returned. *arg* is the C function object.
-
-   ``'c_exception'``
-      A C function has raised an exception.  *arg* is the C function object.
-
    ``'opcode'``
       The interpreter is about to execute a new opcode (see :mod:`dis` for
       opcode details).  The local trace function is called; *arg* is
diff --git a/Misc/NEWS.d/next/Documentation/2018-01-22-21-13-46.bpo-17799.rdZ-Vk.rst b/Misc/NEWS.d/next/Documentation/2018-01-22-21-13-46.bpo-17799.rdZ-Vk.rst
new file mode 100644
index 0000000..ccc52f6
--- /dev/null
+++ b/Misc/NEWS.d/next/Documentation/2018-01-22-21-13-46.bpo-17799.rdZ-Vk.rst
@@ -0,0 +1,2 @@
+Explain real behaviour of sys.settrace and sys.setprofile and their C-API counterparts
+regarding which type of events are received in each function. Patch by Pablo Galindo Salgado.