This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author eric.araujo
Recipients alexis, brian.curtin, eric.araujo, loewis, python-dev, schmir, tarek
Date 2012-03-07.20:27:52
SpamBayes Score 7.1120054e-10
Marked as misclassified No
Message-id <1331152073.36.0.438861822242.issue13719@psf.upfronthosting.co.za>
In-reply-to
Content
I finally got a virtual machine up and running and was able to diagnose the problem.  There are two things.  First, the dist directory (where the msi file will be created) is created relative to the current working directory, which explains why listdir in the temporary project directory caused the test to fail.  Second, after your patch bdist_wininst puts a full path in the dist.dist_files list, but it should use a relative path, like other bdist commands do.  (This choice comes from the fact that setup scripts must be run from their parent directory.)

Fixing the first problem is just a matter of adding os.chdir(project_dir) in the test; the second problem is also easy, using os.path.join and self.dist_dir.  I chose to backout the commits because I won’t be able to make a correct patch right now, I have to configure file sharing between my host and my VM.  Let’s have the buildbots green for a while and then break them again (not :).

Debugging this made me realize that all tests should use os.path.join('dist', filename) instead of e.g. 'dist/blah-1.0.rpm' in test_bdist_rpm; I’ll do that in another commit.
History
Date User Action Args
2012-03-07 20:27:53eric.araujosetrecipients: + eric.araujo, loewis, schmir, tarek, brian.curtin, alexis, python-dev
2012-03-07 20:27:53eric.araujosetmessageid: <1331152073.36.0.438861822242.issue13719@psf.upfronthosting.co.za>
2012-03-07 20:27:52eric.araujolinkissue13719 messages
2012-03-07 20:27:52eric.araujocreate