Ronald Oussoren added the comment:
I'd be in favour of closing this issue, I haven't seen the problem in a while
and it is almost certainly due to a platform bug on OSX 10.6.
--
___
Python tracker rep...@bugs.python.org
Changes by Ned Deily n...@acm.org:
--
resolution: - third party
stage: - resolved
status: open - closed
___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue15398
___
Ned Deily added the comment:
Has anyone seen this problem on other than OS X 10.6? If not, I suggest we
assume it was, in fact, due to a platform bug and close.
--
___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue15398
Ronald Oussoren ronaldousso...@mac.com added the comment:
Whether or not rename is atomic should be irrelevant for test_rename.
Test_rename just performs renames in a tight loop and doesn't do concurrent
access from another thread.
The test result seems to indicate that rename() sometimes
Rafael Caricio raf...@caricio.com added the comment:
I had this problem when I run the tests in my machine (Mac OSX 10.6.8). The
intermittence happen here.
--
nosy: +rafaelcaricio
___
Python tracker rep...@bugs.python.org
Flávio Ribeiro em...@flavioribeiro.com added the comment:
Ronald,
I thought it could be an `atomic` issue by the fact that test_rename transforms
a filename in tmp and then it change again to the original name. Not being
atomic, the rename will not finish the executation of the first one
New submission from Flávio Ribeiro em...@flavioribeiro.com:
Found a intermittent test on UnicodeFileTests.test_rename method.
Python Version: Python 3.3.0b1
Hg commit hash: 3fbfa61634de
MacOS X version 10.6.8
How can be reproduced:
bumblebee:~/dev/cpython[] $ for i in {1..10}; do
Tatiana Al-Chueyr tatiana.alchu...@gmail.com added the comment:
I had the same problem here, after running several times...
$ for i in {1..10}; do ./python.exe -m test test_pep277; done
[1/1] test_pep277
test test_pep277 failed -- Traceback (most recent call last):
File
R. David Murray rdmur...@bitdance.com added the comment:
By the way, you can use -F to run the test suite until it fails:
./python -m test -F test_pep277
--
nosy: +r.david.murray
___
Python tracker rep...@bugs.python.org
R. David Murray rdmur...@bitdance.com added the comment:
Tatiana, are you on a Mac as well?
--
___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue15398
___
Tatiana Al-Chueyr tatiana.alchu...@gmail.com added the comment:
r.david.murray: yes, MacOS X 10.6.8
we managed to reproduce this in other 2 MacBook Pro 8.1.
--
___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue15398
Hugo Lopes Tavares hlt...@gmail.com added the comment:
I had no problems after running for a very long time (using -F). I am using Mac
OSX 10.6.8.
--
nosy: +hltbra
___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue15398
Flávio Ribeiro em...@flavioribeiro.com added the comment:
For doc purposes:
My first shot is that in my machine the os.rename isn't atomic and this could
be generating an overhead somehow. Looking at the code, I found decisions being
made on posixmodule.c about what `rename` function should
Ned Deily n...@acm.org added the comment:
I haven't been able to reproduce this failure on either 10.6 or 10.7 with or
without pydebug and either at 3fbfa61634de or the now slightly more current
tip. What filesystem are you running the test under? What ./configure options
did you use? What
Ned Deily n...@acm.org added the comment:
And OS X doesn't support the *at system calls so HAVE_RENAMEAT will always be
false on OS X. I don't recall every seeing this test fail before.
--
___
Python tracker rep...@bugs.python.org
R. David Murray rdmur...@bitdance.com added the comment:
I found this link:
http://www.weirdnet.nl/apple/rename.html
in issue 8828, which sounds like it is relevant. (I thought I remembered
something about rename not being atomic on OS X.) I didn't read it through,
but maybe Apple fixed
Flávio Ribeiro em...@flavioribeiro.com added the comment:
Hi Ned,
I've used ./configure --with-pydebug make -j2
$ gcc --version
i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5666) (dot 3)
Copyright (C) 2007 Free Software Foundation, Inc.
$ printenv | grep LC_
$ printenv| grep
Ned Deily n...@acm.org added the comment:
Thanks, Flávio. That configuration looks perfectly normal and what I use so
that should not be an issue.
Interesting link, David. I'm not able to reproduce that failure on 10.6.8
using the procedure in the link but I am running 10.6.8 on a virtual
18 matches
Mail list logo
