Exception Chaining missing method for suppressing context

[Benjamin Peterson]

http://bugs.python.org/review/6210/diff/4108/14029
File Lib/test/test_raise.py (right):

http://bugs.python.org/review/6210/diff/4108/14029#newcode87
Lib/test/test_raise.py:87: self.assertTrue(isinstance(e.__context__, TypeError))
assertIsInstance

http://bugs.python.org/review/6210/diff/4108/14031
File Objects/exceptions.c (right):

http://bugs.python.org/review/6210/diff/4108/14031#newcode309
Objects/exceptions.c:309: ;
This is hideous and wrong, since Py_None would still have to be increfed.

http://bugs.python.org/review/6210/diff/4108/14032
File Python/ceval.c (right):

http://bugs.python.org/review/6210/diff/4108/14032#newcode3582
Python/ceval.c:3582: Py_INCREF(Py_None);
There's no point doing this dance, since you're just stealing the reference.

http://bugs.python.org/review/6210/diff/4108/14033
File Python/pythonrun.c (right):

http://bugs.python.org/review/6210/diff/4108/14033#newcode1696
Python/pythonrun.c:1696: ;
Ugly! Figure out a better way to express the conditions.

[eric.araujo]

A few comments on the docs.

http://bugs.python.org/review/6210/diff/4108/14026
File Doc/c-api/exceptions.rst (right):

http://bugs.python.org/review/6210/diff/4108/14026#newcode424
Doc/c-api/exceptions.rst:424: Return the cause (either an exception instance, or
:keyword:`None`,
Should be ``None``, see the Documenting Python section in the devguide.

http://bugs.python.org/review/6210/diff/4108/14026#newcode425
Doc/c-api/exceptions.rst:425: set by ``raise ... from ...``) associated with the
exception as a new
With all the discussion about “from None” and “from ...”, I wonder if this
example would read better as “raise X from Y”.

http://bugs.python.org/review/6210/diff/4108/14026#newcode429
Doc/c-api/exceptions.rst:429: ``__cause__ is Ellipsis``).  If the cause is
:keyword:`None`, the default
I don’t understand this markup.  ``__cause__ is Ellipsis`` looks like an
expression to me, a test; maybe you mean :attr:`cause` is :keyword:`Ellipsis`?

http://bugs.python.org/review/6210/diff/4108/14027
File Doc/library/exceptions.rst (right):

http://bugs.python.org/review/6210/diff/4108/14027#newcode54
Doc/library/exceptions.rst:54: 
The doc for the default exception display code looks clear.  Should the doc for
the traceback module also say something? (and maybe be linked from here)

[amaury.forgeotdarc]

http://bugs.python.org/review/6210/diff/4108/14026
File Doc/c-api/exceptions.rst (right):

http://bugs.python.org/review/6210/diff/4108/14026#newcode425
Doc/c-api/exceptions.rst:425: set by ``raise ... from ...``) associated with the
exception as a new
Isn't "..." another spelling for Ellipsis?  This "raise ... from ..." becomes
confusing.
Maybe use something like "raise anException from anotherException"

http://bugs.python.org/review/6210/diff/4108/14030
File Lib/traceback.py (right):

http://bugs.python.org/review/6210/diff/4108/14030#newcode124
Lib/traceback.py:124: its.append(_iter_chain(cause, False, seen))
Why did you change None to False? this is the custom_tb parameter, which is
either a traceback or None.

http://bugs.python.org/review/6210/diff/4108/14031
File Objects/exceptions.c (right):

http://bugs.python.org/review/6210/diff/4108/14031#newcode309
Objects/exceptions.c:309: ;
Agreed. A Py_INCREF(Py_None) looks more correct here.

http://bugs.python.org/review/6210/diff/4108/14032
File Python/ceval.c (right):

http://bugs.python.org/review/6210/diff/4108/14032#newcode3587
Python/ceval.c:3587: equivalent to 'raise ...' */
Since ... is Ellipsis, this comment is confusing. 
maybe 'raise exc from Ellipsis' and 'raise exc'