Failure on assorted buildbots - Address already in use

A lot of the buildbots are red at the moment, which makes it
harder to 
tell if a checkin broke anything new on other platforms.

I've checked in a change to test_resource that should
hopefully make 
some of the Debian buildbots happier, but several of the
other buildbots 
are reporting a variety of "Address already in
use" errors in the 
subthreads created by test_urllib2.

Anyone have any ideas?

Cheers,
Nick.

-- 
Nick Coghlan   |   ncoghlangmail.com   |   Brisbane,
Australia
------------------------------------------------------------
---
             http://www.boredoma
ndlaziness.org
_______________________________________________
Python-Dev mailing list
Python-Devpython.org
ht
tp://mail.python.org/mailman/listinfo/python-dev
Unsubscribe: http://mail.python.org/mailman/options/p
ython-dev/nessto%40sharedlog.com

2007/7/24, Nick Coghlan <ncoghlangmail.com>:

> some of the Debian buildbots happier, but several of
the other buildbots
> are reporting a variety of "Address already in
use" errors in the
> subthreads created by test_urllib2.

Test pass ok in my machine.

However, if in another terminal I make...

>>> import socket
>>> s = socket.socket()
>>> s.bind(("127.0.0.1", 8080))

...the test fails exactly like in the buildbot.

Maybe the tests should be changed to use a not-so-standard
port.

Regards,

-- 
.    Facundo

Blog: http://www.tanique
til.com.ar/plog/
PyAr: http://www.python.org/ar/
_______________________________________________
Python-Dev mailing list
Python-Devpython.org
ht
tp://mail.python.org/mailman/listinfo/python-dev
Unsubscribe: http://mail.python.org/mailman/options/p
ython-dev/nessto%40sharedlog.com

Facundo Batista wrote:
> 2007/7/24, Nick Coghlan <ncoghlangmail.com>:
> 
> > some of the Debian buildbots happier, but several
of the other buildbots
> > are reporting a variety of "Address already
in use" errors in the
> > subthreads created by test_urllib2.
> 
> Test pass ok in my machine.
> 
> However, if in another terminal I make...
> 
> >>> import socket
> >>> s = socket.socket()
> >>> s.bind(("127.0.0.1", 8080))
> 
> ...the test fails exactly like in the buildbot.
> 
> Maybe the tests should be changed to use a
not-so-standard port.

Or use port 0 to let the operating system pick a free port:

    >>> import socket
    >>> s = socket.socket()
    >>> s.bind(("127.0.0.1", 0))
    >>> s.getsockname()
    ('127.0.0.1', 42669)

-Andrew.

_______________________________________________
Python-Dev mailing list
Python-Devpython.org
ht
tp://mail.python.org/mailman/listinfo/python-dev
Unsubscribe: http://mail.python.org/mailman/options/p
ython-dev/nessto%40sharedlog.com

Andrew Bennetts wrote:
> Facundo Batista wrote:
>> 2007/7/24, Nick Coghlan <ncoghlangmail.com>:
>> Maybe the tests should be changed to use a
not-so-standard port.
> 
> Or use port 0 to let the operating system pick a free
port:
> 
>     >>> import socket
>     >>> s = socket.socket()
>     >>> s.bind(("127.0.0.1", 0))
>     >>> s.getsockname()
>     ('127.0.0.1', 42669)
> 
> -Andrew.

I've changed test_urllib2_localnet to work this way - we'll
see if it 
improves matters.

Cheers,
Nick.

-- 
Nick Coghlan   |   ncoghlangmail.com   |   Brisbane,
Australia
------------------------------------------------------------
---
             http://www.boredoma
ndlaziness.org
_______________________________________________
Python-Dev mailing list
Python-Devpython.org
ht
tp://mail.python.org/mailman/listinfo/python-dev
Unsubscribe: http://mail.python.org/mailman/options/p
ython-dev/nessto%40sharedlog.com

Nick Coghlan wrote:
> Andrew Bennetts wrote:
>> Or use port 0 to let the operating system pick a
free port:
>>
>>     >>> import socket
>>     >>> s = socket.socket()
>>     >>> s.bind(("127.0.0.1",
0))
>>     >>> s.getsockname()
>>     ('127.0.0.1', 42669)
>>
>> -Andrew.
> 
> I've changed test_urllib2_localnet to work this way -
we'll see if it 
> improves matters.

Yep, looks like that did the trick. Facundo, a similar
change may help 
with the GSoC project you're mentoring (the new smtplib
tests failed on 
at least one of the buildbots).

Cheers,
Nick.

-- 
Nick Coghlan   |   ncoghlangmail.com   |   Brisbane,
Australia
------------------------------------------------------------
---
             http://www.boredoma
ndlaziness.org
_______________________________________________
Python-Dev mailing list
Python-Devpython.org
ht
tp://mail.python.org/mailman/listinfo/python-dev
Unsubscribe: http://mail.python.org/mailman/options/p
ython-dev/nessto%40sharedlog.com

2007/7/25, Nick Coghlan <ncoghlangmail.com>:

> Yep, looks like that did the trick. Facundo, a similar
change may help
> with the GSoC project you're mentoring (the new smtplib
tests failed on
> at least one of the buildbots).

Yes! Alan is already working in this (he sent me today a
patch, .

Regards,

-- 
.    Facundo

Blog: http://www.tanique
til.com.ar/plog/
PyAr: http://www.python.org/ar/
_______________________________________________
Python-Dev mailing list
Python-Devpython.org
ht
tp://mail.python.org/mailman/listinfo/python-dev
Unsubscribe: http://mail.python.org/mailman/options/p
ython-dev/nessto%40sharedlog.com