/trunk test_distutils failing on Mac OS X 10.5
Hi all, /trunk test_distutils is failing with the following error on Mac OS X 10.5: Here's the error: test_distutils test test_distutils failed -- Traceback (most recent call last): File "/private/tmp/tmp8UfLPT/python27/Lib/distutils/tests/test_sdist.py", line 342, in test_make_distribution_owner_group self.assertEquals(member.gid, os.getgid()) AssertionError: 0 != 20 It has been a problem for over a week, perhaps longer. I've filed it as: http://bugs.python.org/issue7408 ... So why am I posting this to python-dev? I went to double-check this on the buildbots and noticed that there aren't any Mac OS X buildbots. I would be happy to give one or two people remote access to my Mac OS X 10.5 iMac if someone wanted to set up a buildbot and/or debug this issue further. Tarek, I can give you access immediately through your lyorn account, too. cheers, --titus -- C. Titus Brown, ctb@msu.edu
On Sun, Nov 29, 2009 at 6:32 PM, C. Titus Brown <ctb@msu.edu> wrote:
Hi all,
/trunk test_distutils is failing with the following error on Mac OS X 10.5:
Here's the error:
test_distutils test test_distutils failed -- Traceback (most recent call last): File "/private/tmp/tmp8UfLPT/python27/Lib/distutils/tests/test_sdist.py", line 342, in test_make_distribution_owner_group self.assertEquals(member.gid, os.getgid()) AssertionError: 0 != 20
It has been a problem for over a week, perhaps longer. I've filed it as:
I don't have this problem on my own Mac OS 10.5, I am investigating,
... So why am I posting this to python-dev?
I went to double-check this on the buildbots and noticed that there aren't any Mac OS X buildbots.
I thought we had one at some point,
I would be happy to give one or two people remote access to my Mac OS X 10.5 iMac if someone wanted to set up a buildbot and/or debug this issue further.
Tarek, I can give you access immediately through your lyorn account, too.
Please, yes that would be awesome Thank you
I went to double-check this on the buildbots and noticed that there aren't any Mac OS X buildbots.
That's not true, see http://www.python.org/dev/buildbot/builders/x86 osx.5 trunk Regards, Martin
On Sun, Nov 29, 2009 at 07:28:50PM +0100, "Martin v. L?wis" wrote:
I went to double-check this on the buildbots and noticed that there aren't any Mac OS X buildbots.
That's not true, see
http://www.python.org/dev/buildbot/builders/x86 osx.5 trunk
Ahh! Sorry, was searching for 'mac' somewhere in the string. Those tests are also broken but in different areas from mine. Again, if people would like shell access, just ask. best, --titus -- C. Titus Brown, ctb@msu.edu
On Sun, Nov 29, 2009 at 6:36 PM, C. Titus Brown <ctb@msu.edu> wrote:
On Sun, Nov 29, 2009 at 07:28:50PM +0100, "Martin v. L?wis" wrote:
That's not true, see
http://www.python.org/dev/buildbot/builders/x86 osx.5 trunk
Ahh! Sorry, was searching for 'mac' somewhere in the string.
Those tests are also broken but in different areas from mine. Again, if people would like shell access, just ask.
I think that buildbot is running OS X 10.6 (despite its name). It would definitely be useful to have an OS X 10.5 buildbot as well, since there are significant differences: most obviously, the default build on 10.6 is an LP64 build. I'm not seeing the test_distutils failure you report on my own 10.5 machine, for some reason. Mark
On Sun, Nov 29, 2009 at 06:43:50PM +0000, Mark Dickinson wrote:
On Sun, Nov 29, 2009 at 6:36 PM, C. Titus Brown <ctb@msu.edu> wrote:
On Sun, Nov 29, 2009 at 07:28:50PM +0100, "Martin v. L?wis" wrote:
That's not true, see
http://www.python.org/dev/buildbot/builders/x86 osx.5 trunk
Ahh! ??Sorry, was searching for 'mac' somewhere in the string.
Those tests are also broken but in different areas from mine. ??Again, if people would like shell access, just ask.
I think that buildbot is running OS X 10.6 (despite its name). It would definitely be useful to have an OS X 10.5 buildbot as well, since there are significant differences: most obviously, the default build on 10.6 is an LP64 build.
I'm not seeing the test_distutils failure you report on my own 10.5 machine, for some reason.
Yes, neither is Tarek... I'll see if I can track down why. I don't think I installed anything other than the minimum necessary to build Python, but that's a bit worrisome. cheers, --titus -- C. Titus Brown, ctb@msu.edu
In article <20091129184722.GA32263@idyll.org>, "C. Titus Brown" <ctb@msu.edu> wrote:
On Sun, Nov 29, 2009 at 06:43:50PM +0000, Mark Dickinson wrote:
I'm not seeing the test_distutils failure you report on my own 10.5 machine, for some reason.
Yes, neither is Tarek... I'll see if I can track down why. I don't think I installed anything other than the minimum necessary to build Python, but that's a bit worrisome.
I'm also seeing it on 10.5. Investigating... -- Ned Deily, nad@acm.org
On Sun, Nov 29, 2009 at 8:55 PM, Ned Deily <nad@acm.org> wrote:
In article <20091129184722.GA32263@idyll.org>, "C. Titus Brown" <ctb@msu.edu> wrote:
On Sun, Nov 29, 2009 at 06:43:50PM +0000, Mark Dickinson wrote:
I'm not seeing the test_distutils failure you report on my own 10.5 machine, for some reason.
Yes, neither is Tarek... I'll see if I can track down why. I don't think I installed anything other than the minimum necessary to build Python, but that's a bit worrisome.
I'm also seeing it on 10.5. Investigating...
Following-up in the issue;
Mark Dickinson <dickinsm@gmail.com> writes:
I'm not seeing the test_distutils failure you report on my own 10.5 machine, for some reason.
I don't think it's OSX specific. My FreeBSD slaves (both 6.4 and 7.2) have been getting the same error. It seems related to the sticky bit on /tmp and the fact that the account running the build slave is in the wheel(0) group, so directories/files created beneath /tmp get a group of 0 rather than the user's gid. That affects the temporary files in the distutils test, which is expecting a default of os.getgid(). -- David
participants (6)
-
"Martin v. Löwis" -
C. Titus Brown -
David Bolen -
Mark Dickinson -
Ned Deily -
Tarek Ziadé