Make socket support TIPC.

[albertito]

BPO	1646
Nosy	@gvanrossum, @tiran
Files	0001-Make-socket-support-TIPC.patch 0002-Add-documentation-and-Misc-NEWS-entry-for-TIPC-suppo.patch 0003-Add-unit-tests-for-TIPC-socket-support.patch

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = 'https://github.com/tiran'
closed_at = <Date 2008-01-07.16:19:08.310>
created_at = <Date 2007-12-18.02:54:12.046>
labels = ['type-feature', 'library']
title = 'Make socket support TIPC.'
updated_at = <Date 2008-01-07.16:19:08.308>
user = 'https://bugs.python.org/albertito'

bugs.python.org fields:

activity = <Date 2008-01-07.16:19:08.308>
actor = 'christian.heimes'
assignee = 'christian.heimes'
closed = True
closed_date = <Date 2008-01-07.16:19:08.310>
closer = 'christian.heimes'
components = ['Library (Lib)']
creation = <Date 2007-12-18.02:54:12.046>
creator = 'albertito'
dependencies = []
files = ['9092', '9093', '9094']
hgrepos = []
issue_num = 1646
keywords = ['patch']
message_count = 11.0
messages = ['58721', '58755', '58860', '58885', '58960', '59336', '59378', '59439', '59456', '59458', '59459']
nosy_count = 3.0
nosy_names = ['gvanrossum', 'christian.heimes', 'albertito']
pr_nums = []
priority = 'normal'
resolution = 'fixed'
stage = None
status = 'closed'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue1646'
versions = ['Python 2.6']

[albertito]

TIPC (http://tipc.sf.net) is an open protocol designed for use in
clustered computer environments. It currently has an open source
implementation for Linux (>= 2.6.16), and VxWorks.

The attached patch (against trunk) adds optional Linux-only support for
TIPC in the socket module.

If you want me to make one against 3.0 or any other branch, just let me
know.

Thanks,
Alberto

[gvanrossum]

I'm okay with adding this, it doesn't add much code and is properly
safeguarded by #ifdefs and has a configure.in patch.

I haven't reviewed the code though.

[albertito]

On Tue, Dec 18, 2007 at 07:10:39PM -0000, Guido van Rossum wrote:

Guido van Rossum added the comment:

I'm okay with adding this, it doesn't add much code and is properly
safeguarded by #ifdefs and has a configure.in patch.

Thanks!

Is there anything else I need to do in order to get this ready for
inclusion?

Thanks again,
Alberto

[gvanrossum]

Tickle the interest of one of the many folks with commit privileges.
Since 2.6 isn't going to be released for months there's no great hurry,
but waiting until the last minute would be a mistake.

[tiran]

I've chatted with albertito a few days ago and adviced him to post the
patch here. I'll take it.

From the first review the patch looks good. The code is written cleanly
and it uses the appropriate API. The new code is surrounded by ifdefs
and it has a patch for configure.in. I'll do the autoreconf to update
the other files locally.

The patch is missing documentation updates, Misc/NEWS update and some
unit tests. The docs should list the new constants and explain TIPC
briefly in a few sentences. You don't need to write a novell ;)

[tiran]

*PING*

I'm still waiting for a new patch. Your feature patch has been accepted
but I won't commit it without doc updates.

[albertito]

On Sat, Jan 05, 2008 at 10:08:15PM -0000, Christian Heimes wrote:

I'm still waiting for a new patch. Your feature patch has been accepted
but I won't commit it without doc updates.

I know, but holidays were not the best time to write documentation.

I'll probably send you the updated patch sometime this week; thanks for
the ping =)

Thanks,
Alberto

[albertito]

On Sun, Jan 06, 2008 at 02:45:35PM -0000, Alberto Bertogli wrote:

I'll probably send you the updated patch sometime this week; thanks for
the ping =)

Here are the three patches, rebased to the SVN commit 59815.

The first one is the same I've already submitted, the second adds the
documentation and the third two testcases.

The testcases are simple, but I think cover a reasonable amount of TIPC
specific code. If you want more, let me know.

The documentation is simple, as you told me, just a paragraph explaining
about the address format and another explaining the constants. Again, if
you want me to change anything, just let me know.

Finally, if you want me to rebase this on top of any other branch, you
guessed it, just let me know ;)

Thanks a lot,
Alberto

[tiran]

The unit tests don't work for me (Ubuntu 7.10, Linux 2.6.22, i386)

>>> srv = socket.socket(socket.AF_TIPC, socket.SOCK_RDM)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/heimes/dev/python/trunk/Lib/socket.py", line 177, in __init__
    _sock = _realsocket(family, type, proto)
socket.error: [Errno 97] Address family not supported by protocol

Do I have to set up my network device for TIPC?

[albertito]

On Mon, Jan 07, 2008 at 03:37:54PM -0000, Christian Heimes wrote:
> The unit tests don't work for me (Ubuntu 7.10, Linux 2.6.22, i386)
> 
> >>> srv = socket.socket(socket.AF_TIPC, socket.SOCK_RDM)
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
>   File "/home/heimes/dev/python/trunk/Lib/socket.py", line 177, in __init__
>     _sock = _realsocket(family, type, proto)
> socket.error: [Errno 97] Address family not supported by protocol
> 
> Do I have to set up my network device for TIPC?

No, by default you have a single-node cluster fully capable of TIPC
networking. But you do need the TIPC module loaded; have you tried
modprobe tipc?

If that works, the test cases should pass.

I understand that this might be a problem because there is no time to
check at runtime if you have TIPC enabled in your kernel (besides the
obvious way of trying to create the socket and see if it fails); but you
have the same problem with any network protocol (although obviously is
much more common to have TIPC disabled than the rest of the supported
ones).

Do you want me to check for this in the conditional for inclusion?

Thanks,
Alberto

[tiran]

Committed in r59819