[issue22438] eventlet broke by python 2.7.x

[Nick Coghlan]

Nick Coghlan added the comment:

Just noting that if we decided to reconsider addressing this upstream for the 
benefit of folks backporting PEP 493, Robert Kuska wrote a patch to restore 
sslwrap for the RHEL/CentOS PEP 466 backport:

* 
https://git.centos.org/blob/rpms!python.git/f63228654ecef84a78c552dac832f4cd939cf584/SPECS!python.spec#L989
* 
https://git.centos.org/blob/rpms!python.git/f63228654ecef84a78c552dac832f4cd939cf584/SOURCES!00221-pep466-backport-sslwrap-c-ssl.patch

--
nosy: +ncoghlan, rkuska

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Denis Bilenko]

Denis Bilenko added the comment:

gevent's ssl support is also broken by 2.7.9.

https://github.com/gevent/gevent/issues/477

IMO, it is totally unexpected to have some API (even though it's undocumented 
and internal) removed in non-major release.

Even though both gevent and eventlet can be fixed, there still be combinations 
of versions that break (python = 2.7.9  gevent = 1.0.1)

Please put _ssl.sslwrap back. It would save a lot of people a lot of time. I 
don't mind fixing gevent not to use it, but there's nothing I can do about 
versions already released.

--
nosy: +Denis.Bilenko

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Arfrever Frehtes Taifersar Arahesis]

Arfrever Frehtes Taifersar Arahesis added the comment:

Private API is expected to be allowed to be deleted or incompatibly changed in 
any micro release.

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Marc-Andre Lemburg]

Marc-Andre Lemburg added the comment:

Even though it may have been considered a private API (*), users certainly 
won't understand that their application just broke because of a Python patch 
level release upgrade, so if possible, I think the API should be added back and 
flagged as private, but used by at least eventlet and gevent.

Alternatively: Why not add it back and make it an officially documented API ?

(*) The API is missing the leading underscore, so it's not really private by 
our usual conventions, it's just undocumented.

--
nosy: +lemburg

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Arfrever Frehtes Taifersar Arahesis]

Arfrever Frehtes Taifersar Arahesis added the comment:

_ssl has leading underscore.
Privateness is inherited, so both A._B.C and A._B._D are private.

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Marc-Andre Lemburg]

Marc-Andre Lemburg added the comment:

On 07.11.2014 11:12, Arfrever Frehtes Taifersar Arahesis wrote:
 
 Arfrever Frehtes Taifersar Arahesis added the comment:
 
 _ssl has leading underscore.
 Privateness is inherited, so both A._B.C and A._B._D are private.

No, the use of the underscore in _ssl is per convention that C
implementation part of stdlib modules are moved into modules that
start with an underscore. This doesn't mean that the APIs in
those modules are private, otherwise many C implementations we have
in the stdlib would be private :-)

Also note that _ssl.sslwrap is special in that it's the main
interface between _ssl and ssl.

BTW: The sslwrap_simple() API was also removed in 2.7.9.

Note: Any libraries that need to monkey patch the Python
network stdlib will need access to these APIs. Given that the
ssl implementation changed a lot in 2.7.9, I think special care
has to be taken not to break too many of these. Using gevent
and eventlet as test cases for whether backwards compatibility
is good enough sounds like a workable approach, IMO.

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Arfrever Frehtes Taifersar Arahesis]

Arfrever Frehtes Taifersar Arahesis added the comment:

 No, the use of the underscore in _ssl is per convention that C
 implementation part of stdlib modules are moved into modules that
 start with an underscore. This doesn't mean that the APIs in
 those modules are private, otherwise many C implementations we have
 in the stdlib would be private :-)

The non-private C-implemented modules are these:

$ cd /usr/lib64/python2.7/lib-dynload ; echo [^_]*.so
array.so audioop.so binascii.so bz2.so cmath.so cPickle.so crypt.so 
cStringIO.so datetime.so dbm.so fcntl.so future_builtins.so gdbm.so grp.so 
itertools.so linuxaudiodev.so math.so mmap.so nis.so operator.so ossaudiodev.so 
parser.so pyexpat.so readline.so resource.so select.so spwd.so strop.so 
syslog.so termios.so time.so unicodedata.so zlib.so

_[^_]-prefixed, undocumented modules (amongst whom are both _[^_].py and 
_[^_].so) should be treated as private modules for usage only by public modules 
in standard library.

(_winreg is the only _[^_]-prefixed, documented module in CPython 2.7.)

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Marc-Andre Lemburg]

Marc-Andre Lemburg added the comment:

On 07.11.2014 11:52, Arfrever Frehtes Taifersar Arahesis wrote:
 
 Arfrever Frehtes Taifersar Arahesis added the comment:
 
 No, the use of the underscore in _ssl is per convention that C
 implementation part of stdlib modules are moved into modules that
 start with an underscore. This doesn't mean that the APIs in
 those modules are private, otherwise many C implementations we have
 in the stdlib would be private :-)
 
 The non-private C-implemented modules are these:
 
 $ cd /usr/lib64/python2.7/lib-dynload ; echo [^_]*.so
 array.so audioop.so binascii.so bz2.so cmath.so cPickle.so crypt.so 
 cStringIO.so datetime.so dbm.so fcntl.so future_builtins.so gdbm.so grp.so 
 itertools.so linuxaudiodev.so math.so mmap.so nis.so operator.so 
 ossaudiodev.so parser.so pyexpat.so readline.so resource.so select.so spwd.so 
 strop.so syslog.so termios.so time.so unicodedata.so zlib.so
 
 _[^_]-prefixed, undocumented modules (amongst whom are both _[^_].py and 
 _[^_].so) should be treated as private modules for usage only by public 
 modules in standard library.
 
 (_winreg is the only _[^_]-prefixed, documented module in CPython 2.7.)

I think you're misunderstanding: just because an API is implemented
in one of the _module.c modules, doesn't imply that those APIs are
private. We have for a long time now used the approach to have a Python
module as main entry point and the _module.c modules providing the
C level bits needed by the Python module.

Regardless of whether the API can be considered private or not, packages
which have to monkey patch the low-level APIs in the network modules
will need access to those APIs in order to tap into the layers and
the sslwrap() function was/is such a low level API.

Also note that due to the major implementation changes in the ssl
module for 2.7.9 such quirks are actually expected and so it's good
that we are getting reports from eventlet and gevent on these
issues. After all, what use is a safer ssl module if your application
doesn't work anymore ;-)

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Arfrever Frehtes Taifersar Arahesis]

Arfrever Frehtes Taifersar Arahesis added the comment:

Monkey-patching is as supported as using private API.
Maintainers of third-party projects monkey-patching something or using private 
API should expect to sporadically have to adjust their projects to new Python 
versions.

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Marc-Andre Lemburg]

Marc-Andre Lemburg added the comment:

On 07.11.2014 11:30, Marc-Andre Lemburg wrote:
 BTW: The sslwrap_simple() API was also removed in 2.7.9.

Scratch that. I was in the wrong work dir :-)

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Marc-Andre Lemburg]

Marc-Andre Lemburg added the comment:

On 07.11.2014 12:49, Marc-Andre Lemburg wrote:
 BTW: The sslwrap_simple() API was also removed in 2.7.9.
 
 Scratch that. I was in the wrong work dir :-)

Hmm, even though the API is still there, it uses _ssl.sslwrap() as
well, so it won't work anymore either.

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Arfrever Frehtes Taifersar Arahesis]

Arfrever Frehtes Taifersar Arahesis added the comment:

 Hmm, even though the API is still there, it uses _ssl.sslwrap() as
 well, so it won't work anymore either.

It was fixed in bug #22523.

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Antoine Pitrou]

Antoine Pitrou added the comment:

It's not a mere matter of putting back the code... The 3.x ssl implementation 
which was backported uses a slightly different approach from the 2.x 
implementation, so it's not obvious we can recreate an entirely compatible 
implementation of_ssl.sslwrap().

As a matter of fact, gevent's fix uses some frame locals hackery to lookup the 
caller's self variable, which means it probably won't work in the general 
case:
https://github.com/Eugeny/ajenti/commit/54442ccb2b9ee24af15500557e7dd7b2f58acb97

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Marc-Andre Lemburg]

Marc-Andre Lemburg added the comment:

On 07.11.2014 13:12, Antoine Pitrou wrote:
 
 It's not a mere matter of putting back the code... The 3.x ssl implementation 
 which was backported uses a slightly different approach from the 2.x 
 implementation, so it's not obvious we can recreate an entirely compatible 
 implementation of_ssl.sslwrap().
 
 As a matter of fact, gevent's fix uses some frame locals hackery to lookup 
 the caller's self variable, which means it probably won't work in the 
 general case:
 https://github.com/Eugeny/ajenti/commit/54442ccb2b9ee24af15500557e7dd7b2f58acb97

Yes, that hack will probably only work for gevent.

Is there a reason why caller_self needs to be passed to
context._wrap_socket() ?

I can't even find the ssl_sock kw args used in the hack in the current
2.7.9 code. The method only has a server_name argument.

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Marc-Andre Lemburg]

Marc-Andre Lemburg added the comment:

Looking at recent comments on the gevent ticket, the 2.7.9
changes are already causing problems for people since apparently
the changes were backported to 2.7.8 by some vendors.

https://github.com/gevent/gevent/issues/477

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Alex Gaynor]

Alex Gaynor added the comment:

FWIW, that code is all significantly simplified by the patch in 
http://bugs.python.org/issue22559

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Benjamin Peterson]

Benjamin Peterson added the comment:

Definitely a case of eventlet playing with fire (private apis).

--
resolution:  - not a bug
status: open - closed

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Alex Gaynor]

New submission from Alex Gaynor:

https://github.com/eventlet/eventlet/issues/135

--
components: Library (Lib)
messages: 227067
nosy: alex, benjamin.peterson, christian.heimes, dstufft, giampaolo.rodola, 
janssen, pitrou
priority: normal
severity: normal
status: open
title: eventlet broke by python 2.7.x
versions: Python 2.7

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Alex Gaynor]

Alex Gaynor added the comment:

It looks like something was removed from the `_ssl` module; is that considered 
an implementation detail, or does it need to be added back?

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Alex Gaynor]

Alex Gaynor added the comment:

(It didn't have any direct tests as far as I can tell, which is why I didn't 
catch the regression)

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Antoine Pitrou]

Antoine Pitrou added the comment:

Certainly not an official API anyway. Let eventlet get their stuff straight.

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Barry A. Warsaw]

Changes by Barry A. Warsaw ba...@python.org:


--
nosy: +barry

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Barry A. Warsaw]

Barry A. Warsaw added the comment:

I tend to agree.  I don't even think it was documented.  I wonder though if it 
makes sense to at least mention this in the PEP and/or release notes for 2.7.9.

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue22438] eventlet broke by python 2.7.x

[Arfrever Frehtes Taifersar Arahesis]

Changes by Arfrever Frehtes Taifersar Arahesis arfrever@gmail.com:


--
nosy: +Arfrever

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue22438
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com