Éric Araujo mer...@netwok.org added the comment:
Bumping up to release blocker, as I don’t want the next bugfixes to go out
without reverting the last changeset (but won’t right know as I need to fix the
bug there was on Windows).
--
nosy: +benjamin.peterson, georg.brandl
priority:
Nadeem Vawda nadeem.va...@gmail.com added the comment:
I'm no longer getting the failures on either Ubuntu or Windows (and the
Windows buildbots are now green), so tentatively marking this as fixed.
Feel free to reopen if something is still broken.
--
resolution: - fixed
stage: needs
Éric Araujo mer...@netwok.org added the comment:
This problem was not trivial to find, because it appears that test execution
order may not be
entirely deterministic: I couldn't see any other reason why the flag would
have different values
on different machines.
On my machine, it looks
Antoine Pitrou pit...@free.fr added the comment:
Antoine, I appreciate that you took time to fix this bug while I was
without Internet and without Windows, but unfortunately I will have to
backout your commit. Postel’s Law doesn’t win here: It is documented
that the MANIFEST template only
Éric Araujo mer...@netwok.org added the comment:
That's should, not must.
I read that “should” as a polite “must”.
Also, I thought people did undocumented things with distutils, and we
had to support these undocumented uses?
People rely on undocumented features and sometimes on bugs. Thus,
Antoine Pitrou pit...@free.fr added the comment:
A commit adding problem should be fixed or reverted
The point is that fixing it may take tome. Reverting is fine by me.
But we have no way of knowing you will be taking tome to do it.
Ideally, you should have reverted it yourself (or applied
Roundup Robot devn...@psf.upfronthosting.co.za added the comment:
New changeset 0a94e2f807c7 by Antoine Pitrou in branch '3.2':
Issue #13193: fix distutils.filelist.FileList under Windows
http://hg.python.org/cpython/rev/0a94e2f807c7
New changeset 80d5040f2a78 by Antoine Pitrou in branch '3.2':
Vinay Sajip vinay_sa...@yahoo.co.uk added the comment:
Re. Paul Moore's comment - IMO he's right about the problem, but changing only
packaging.manifest._translate_pattern doesn't do it. The equivalent fix has to
be made in distutils.filelist.translate_pattern. I've made the change in the
Éric Araujo mer...@netwok.org added the comment:
[global variables]
one possible approach might be: Have those bindings be instance variables in
a Database class in
database.py, and have a module-level binding to an instance of it. Then,
tests can have their
own instance which is thrown
Vinay Sajip vinay_sa...@yahoo.co.uk added the comment:
entirely deterministic: I couldn't see any other reason why the flag
would have different values
on different machines.
Sorry, what flag?
By flag I mean _cache_generated_egg (flag as in Boolean value)
Patches for upstream cpython
Éric Araujo mer...@netwok.org added the comment:
By flag I mean _cache_generated_egg (flag as in Boolean value)
Ah, I had forgotten this earlier message:
I get the opposite failure to Nadeem as far as
InstallDataTestCase.test_resources: it works on
Ubuntu 64-bit, but fails on 32-bit.
Vinay Sajip vinay_sa...@yahoo.co.uk added the comment:
That it depends on the architecture currently baffles me.
The only explanation I can come up with is that on different machines, the
order of the tests might be slightly different. That would allow the flag to be
set differently on
Vinay Sajip vinay_sa...@yahoo.co.uk added the comment:
I get the opposite failure to Nadeem as far as
InstallDataTestCase.test_resources: it works on Ubuntu 64-bit, but fails on
32-bit. Digging into it a bit further, I find that _generate_cache in
Lib/packaging/database.py returns prematurely
Vinay Sajip vinay_sa...@yahoo.co.uk added the comment:
I added the packaging.database.clear_cache() at the end of setUp(); that works,
too.
Gotta love global variables - not! ;-)
--
___
Python tracker rep...@bugs.python.org
Vinay Sajip vinay_sa...@yahoo.co.uk added the comment:
I looked in packaging/test, and noticed that database.get_distribution is
called from test_depgraph, test_xmlrpc and test_database. These may also need
looking over to see if cache invalidation needs to happen.
--
Nadeem Vawda nadeem.va...@gmail.com added the comment:
I added the packaging.database.clear_cache() at the end of setUp(); that
works, too.
That fixes the InstallDataTestCase failure for me as well
(on both Ubuntu 11.10 64-bit and Windows 7 64-bit).
I get the opposite failure to Nadeem as
Roumen Petrov bugtr...@roumenpetrov.info added the comment:
Test test_install and test_command_install_data interference cache . Adding
clear_cache as clean up routine will resolve reported issue with Spamlib. The
patch is trivial. Also one of the test set clear_cache as cleanup.
Most
Éric Araujo mer...@netwok.org added the comment:
Thanks, I’ll add the disable_cache call.
Westley: I’ve installed Arch. Is there a command I can run to get zlib,
openssl, gcc, make and all that, similar to “aptitude build-dep python3.2”?
--
___
Éric Araujo mer...@netwok.org added the comment:
I’ll open another reports for the possible bug in _generate_cache and for
review of the other tests calling get_distribution.
Test test_install and test_command_install_data interference cache
I’ve added checks in regrtest to make sure that
Roundup Robot devn...@psf.upfronthosting.co.za added the comment:
New changeset 4eee9dd61147 by Éric Araujo in branch 'default':
Try to fix buildbot failures from #13193
http://hg.python.org/cpython/rev/4eee9dd61147
--
nosy: +python-dev
___
Python
Vinay Sajip vinay_sa...@yahoo.co.uk added the comment:
[citation needed]. First, mutable objects with a global name (let’s not use
variable for Python) are not unconditionally evil; see sys.path and
sys.modules for example. Second, how concretely would you change that
implementation? We
Vinay Sajip vinay_sa...@yahoo.co.uk added the comment:
FTR, I haven't checked this on Ubuntu 32-bit. It's rather odd for the
failure to be architecture-dependent. Are these two machines running
different versions of Ubuntu, perhaps?
They are (Ubuntu Natty 32-bit failing, Ubuntu Oneiric
Nadeem Vawda nadeem.va...@gmail.com added the comment:
How strange. I've had no problem reproducing the failures; they've
occurred every time I've run test_packaging since the bug was opened
(on both Ubuntu and Windows). And the Windows buildbots seem to be
failing consistently as well...
Vinay Sajip vinay_sa...@yahoo.co.uk added the comment:
How strange. I've had no problem reproducing the failures; they've
occurred every time I've run test_packaging since the bug was opened
(on both Ubuntu and Windows). And the Windows buildbots seem to be
failing consistently as well...
Éric Araujo mer...@netwok.org added the comment:
anikom15, can you give me a URI to download the OS you used? I don’t know when
I’ll be able to download Windows, but surely a linux-using OS will take less
space and time to download and install in a VM.
--
Nadeem Vawda nadeem.va...@gmail.com added the comment:
I've been able to consistently reproduce the test_resources failure on Ubuntu
11.10 64-bit, FWIW.
--
___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue13193
Westley Martínez aniko...@gmail.com added the comment:
http://www.archlinux.org/download/
It's a minimalist distribution but if you read through the install guide or
beginner's guide you'll be fine.
--
___
Python tracker rep...@bugs.python.org
Éric Araujo mer...@netwok.org added the comment:
Spamlib-0.1.dist-info is in a directory on sys.path, so the error makes no
sense to me. Can you inspect the content of that directory? (i.e os.listdir
and file contents) Also, if you can interrupt the test to get a Python or pdb
shell, could
Francisco Martín Brugué franci...@email.de added the comment:
Hi Eli,
I cannot find the file/Distro:
find / -name *'dist-info' 2 /dev/null
/home/ci/cpython/Lib/packaging/tests/fake_dists/grammar-1.0a4.dist-info
/home/ci/cpython/Lib/packaging/tests/fake_dists/towel_stuff-0.1.dist-info
Francisco Martín Brugué franci...@email.de added the comment:
The changeset was:
$ hg tip
changeset: 73075:d4839fea4a5a
[...]
--
___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue13193
___
Nadeem Vawda nadeem.va...@gmail.com added the comment:
Spamlib-0.1.dist-info is in a directory on sys.path, so the error makes no
sense to me. Can you inspect the content of that directory? (i.e os.listdir
and file contents) Also, if you can interrupt the test to get a Python or
pdb
Francisco Martín Brugué franci...@email.de added the comment:
Where should be the distro?
You won't find it in the hg repository - it gets created by the test, and
then deleted afterwards.
Thanks for the info
The changeset was:
$ hg tip
changeset: 73075:d4839fea4a5a
[...]
That's
Nadeem Vawda nadeem.va...@gmail.com added the comment:
Issue 13230 (closed as duplicate) reports another Linux failure in
test_resources.
Changing the title, since it isn't Windows-specific.
--
nosy: +anikom15
stage: - needs patch
___
Python
Changes by Nadeem Vawda nadeem.va...@gmail.com:
--
title: test_packaging and test_distutils failures under Windows -
test_packaging and test_distutils failures
___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue13193
Nadeem Vawda nadeem.va...@gmail.com added the comment:
The test creates a temporary directory which is inserted at the head of
sys.path.
packaging.database.get_distribution should thus find Spamlib-0.1.dist-info.
Can someone with a Windows install help me with this? Printing sys.path and
Changes by Florent Xicluna florent.xicl...@gmail.com:
--
nosy: +flox
___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue13193
___
___
Python-bugs-list
Antoine Pitrou pit...@free.fr added the comment:
Similar test failure under x86 OpenIndiana:
==
ERROR: test_resources
(packaging.tests.test_command_install_data.InstallDataTestCase)
Éric Araujo mer...@netwok.org added the comment:
The first failure looks like a bug in the manifest recursive-include code:
FAIL: test_process_template (distutils.tests.test_filelist.FileListTestCase)
--
Traceback (most recent
Paul Moore p.f.mo...@gmail.com added the comment:
I'm not getting the second error on my home PC. As the failing buildbot is
mine, I'll have a look on there to see if I can reproduce when I get the chance.
--
___
Python tracker
Paul Moore p.f.mo...@gmail.com added the comment:
First one - the problem is in packaging.manifest._translate_pattern, which uses
os.path.join on regex parts. That won't work on Windows where os.sep is a
backslash, as the backslash is a RE metacharacter.
Actually, the file list seems to only
New submission from Antoine Pitrou pit...@free.fr:
http://www.python.org/dev/buildbot/all/builders/x86%20XP-5%203.x/builds/3593/steps/test/logs/stdio
--
assignee: tarek
components: Distutils, Distutils2, Library (Lib), Tests
messages: 145643
nosy: alexis, eric.araujo, pitrou, tarek
41 matches
Mail list logo