Index: Doc/library/doctest.rst
===================================================================
--- Doc/library/doctest.rst	(revision 76753)
+++ Doc/library/doctest.rst	(working copy)
@@ -534,10 +534,20 @@
    example, an example expecting ``ValueError: 42`` will pass if the actual
    exception raised is ``ValueError: 3*14``, but will fail, e.g., if
    :exc:`TypeError` is raised.
-
-   Note that a similar effect can be obtained using :const:`ELLIPSIS`, and
-   :const:`IGNORE_EXCEPTION_DETAIL` may go away when Python releases prior to 2.4
-   become uninteresting.  Until then, :const:`IGNORE_EXCEPTION_DETAIL` is the only
+   
+   It will also ignore the module name used in Python 3 doctest reports. Hence
+   both these variations will work:
+   
+      >>> raise ValueError('message') #doctest: +IGNORE_EXCEPTION_DETAIL
+      Traceback (most recent call last):
+      ValueError: message
+      
+      >>> raise ValueError('message') #doctest: +IGNORE_EXCEPTION_DETAIL
+      Traceback (most recent call last):
+      builtin.ValueError: message
+      
+   Note that a similar effect can be obtained using :const:`ELLIPSIS`, but
+   using :const:`IGNORE_EXCEPTION_DETAIL` is often clearer. It is also the only
    clear way to write a doctest that doesn't care about the exception detail yet
    continues to pass under Python releases prior to 2.4 (doctest directives appear
    to be comments to them).  For example, ::
Index: Lib/doctest.py
===================================================================
--- Lib/doctest.py	(revision 76753)
+++ Lib/doctest.py	(working copy)
@@ -1276,9 +1276,9 @@
 
                 # Another chance if they didn't care about the detail.
                 elif self.optionflags & IGNORE_EXCEPTION_DETAIL:
-                    m1 = re.match(r'[^:]*:', example.exc_msg)
-                    m2 = re.match(r'[^:]*:', exc_msg)
-                    if m1 and m2 and check(m1.group(0), m2.group(0),
+                    m1 = re.match(r'(?:[^:]*\.)?([^:]*:)', example.exc_msg)
+                    m2 = re.match(r'(?:[^:]*\.)?([^:]*:)', exc_msg)
+                    if m1 and m2 and check(m1.group(1), m2.group(1),
                                            self.optionflags):
                         outcome = SUCCESS
 
Index: Lib/test/test_doctest.py
===================================================================
--- Lib/test/test_doctest.py	(revision 76753)
+++ Lib/test/test_doctest.py	(working copy)
@@ -863,7 +863,78 @@
     >>> test = doctest.DocTestFinder().find(f)[0]
     >>> doctest.DocTestRunner(verbose=False).run(test)
     TestResults(failed=0, attempted=1)
+    
+IGNORE_EXCEPTION_DETAIL also ignores difference in exception formatting 
+between Python versions. For example, in Python 3.x, the module path of
+the exception is in the output:
 
+    >>> def f(x):
+    ...     r'''
+    ...     >>> from http.client import HTTPException
+    ...     >>> raise HTTPException('message')
+    ...     Traceback (most recent call last):
+    ...     http.client.HTTPException: message
+    ...     '''
+    >>> test = doctest.DocTestFinder().find(f)[0]
+    >>> doctest.DocTestRunner(verbose=False).run(test)
+    TestResults(failed=0, attempted=2)
+
+But in Python 2 the module path is not included, an therefore a test must look
+like the following test to succeed in Python 2. But that test will fail under
+Python 3.
+    
+    >>> def f(x):
+    ...     r'''
+    ...     >>> from http.client import HTTPException
+    ...     >>> raise HTTPException('message')
+    ...     Traceback (most recent call last):
+    ...     HTTPException: message
+    ...     '''
+    >>> test = doctest.DocTestFinder().find(f)[0]
+    >>> doctest.DocTestRunner(verbose=False).run(test)
+    ... # doctest: +ELLIPSIS
+    **********************************************************************
+    File ..., line 4, in f
+    Failed example:
+        raise HTTPException('message')
+    Expected:
+        Traceback (most recent call last):
+        HTTPException: message
+    Got:
+        Traceback (most recent call last):
+        ...
+        http.client.HTTPException: message
+    TestResults(failed=1, attempted=2)
+    
+However, with IGNORE_EXCEPTION_DETAIL, the module name of the exception 
+(if any) will be ignored:
+
+    >>> def f(x):
+    ...     r'''
+    ...     >>> from http.client import HTTPException
+    ...     >>> raise HTTPException('message') #doctest: +IGNORE_EXCEPTION_DETAIL
+    ...     Traceback (most recent call last):
+    ...     HTTPException: message
+    ...     '''
+    >>> test = doctest.DocTestFinder().find(f)[0]
+    >>> doctest.DocTestRunner(verbose=False).run(test)
+    TestResults(failed=0, attempted=2)
+
+The module path will be completely ignored, so two different module paths will
+still pass if IGNORE_EXCEPTION_DETAIL is given. This is intentional, so it can
+be used when exceptions have changed module.
+
+    >>> def f(x):
+    ...     r'''
+    ...     >>> from http.client import HTTPException
+    ...     >>> raise HTTPException('message') #doctest: +IGNORE_EXCEPTION_DETAIL
+    ...     Traceback (most recent call last):
+    ...     foo.bar.HTTPException: message
+    ...     '''
+    >>> test = doctest.DocTestFinder().find(f)[0]
+    >>> doctest.DocTestRunner(verbose=False).run(test)
+    TestResults(failed=0, attempted=2)
+
 But IGNORE_EXCEPTION_DETAIL does not allow a mismatch in the exception type:
 
     >>> def f(x):