Created on 2010-06-29 14:27 by csernazs, last changed 2011-01-03 17:59 by eric.araujo. This issue is now closed.
| Files | ||||
|---|---|---|---|---|
| File name | Uploaded | Description | Edit | |
| test_site.py.diff | csernazs, 2010-06-29 14:27 | Fix for test_site.py | review | |
| fix-9115.diff | eric.araujo, 2010-12-26 03:14 | |||
| Messages (4) | |||
|---|---|---|---|
| msg108910 - (view) | Author: Zsolt Cserna (csernazs) | Date: 2010-06-29 14:27 | |
On systems where there's no unsetenv function (for example solaris 8), test_site fails because the PYTHONUSERBASE environment variable remains set to "xoxo" (set in the previous tests). This results a failed test_s_option test. |
|||
| msg124663 - (view) | Author: Éric Araujo (eric.araujo) *  |
Date: 2010-12-26 03:14 | |
Attached patch uses a copy of os.environ, like other tests already do. Can you test it? Also, is the lack of unsetenv a problem only for test_s_option? If not, we should fix the test infrastructure, not each test. |
|||
| msg125188 - (view) | Author: Zsolt Cserna (csernazs) | Date: 2011-01-03 16:14 | |
I confirm that this patch fixes the problem. Thanks. On my systems I haven't seen other bugs related to unsetenv - however, it might be useful to fix subprocess.Popen and subprocess.call to use the os.environ by default (but this would be another request or discussion). |
|||
| msg125207 - (view) | Author: Éric Araujo (eric.araujo) * |
Date: 2011-01-03 17:59 | |
Fixed in r87691 (py3k), r87693 (3.1) and r87694 (2.7), thanks! |
|||
| History | |||
|---|---|---|---|
| Date | User | Action | Args |
| 2011-01-03 17:59:02 | eric.araujo | set | status: open -> closed messages: + msg125207 resolution: fixed stage: patch review -> committed/rejected |
| 2011-01-03 16:14:39 | csernazs | set | messages: + msg125188 |
| 2010-12-26 03:14:40 | eric.araujo | set | files:
+ fix-9115.diff versions: + Python 3.1, Python 3.2 nosy: + eric.araujo messages: + msg124663 stage: patch review |
| 2010-06-29 14:27:57 | csernazs | create | |