Created on 2013-01-08 21:39 by zach.ware, last changed 2013-01-10 02:34 by ezio.melotti. This issue is now closed.
| Files | ||||
|---|---|---|---|---|
| File name | Uploaded | Description | Edit | |
| test_bisect_fix.diff | zach.ware, 2013-01-08 21:39 | test_bisect.py fix, version 1 | review | |
| test_bisect_fix.v2.diff | zach.ware, 2013-01-09 15:33 | Version 2 | review | |
| Messages (7) | |||
|---|---|---|---|
| msg179374 - (view) | Author: Zachary Ware (zach.ware) *  |
Date: 2013-01-08 21:39 | |
Here's the fix for test_bisect.py. This one has an extra change, that of replacing the manual method of obtaining c_bisect and py_bisect with two calls to support.import_fresh_module. This one also has an extra wrench thrown in the gears by including a few doctests. I haven't changed anything in relation to those doctests, but I wonder if there's really any point to keeping them there considering they claim to be pulled from the docs but are in fact significantly different from what is in the docs these days. |
|||
| msg179413 - (view) | Author: Ezio Melotti (ezio.melotti) * |
Date: 2013-01-09 06:17 | |
The best thing would be converting them to unittest (if there aren't already unittests for that). A general comment about these patches: I think it would be better to remove test_main() (and related unnecessary code) with unittest.main(). One of the advantage of doing it now is that you can simply test discovery by doing ./python Lib/test/test_foo.py. This is currently being discussed on #16748, but I think we pretty much agree on this. |
|||
| msg179416 - (view) | Author: Zachary Ware (zach.ware) * |
Date: 2013-01-09 06:47 | |
> A general comment about these patches: I think it would be better to remove test_main() (and related unnecessary code) with unittest.main(). One of the advantage of doing it now is that you can simply test discovery by doing ./python Lib/test/test_foo.py. > > This is currently being discussed on #16748, but I think we pretty much agree on this. > Ah yes, I meant to do that and completely forgot about it. I'll go through and do that on the ones I've made patches for so far in the morning. |
|||
| msg179417 - (view) | Author: Ezio Melotti (ezio.melotti) * |
Date: 2013-01-09 07:03 | |
I think that here all the test_main() can be removed except support.run_doctest(test_bisect, verbose). This should probably be moved after unittest.main() (unless the doctests are removed/converted to unittest). Refcounting is already provided by regrtest. |
|||
| msg179454 - (view) | Author: Zachary Ware (zach.ware) * |
Date: 2013-01-09 15:33 | |
Here's version 2 of the patch, which converts the doctests to a standard testcase and removes test_main(). |
|||
| msg179512 - (view) | Author: Roundup Robot (python-dev) | Date: 2013-01-10 02:33 | |
New changeset f32cb5389a4b by Ezio Melotti in branch '3.3': #16897: test_bisect now works with unittest test discovery. Initial patch by Zachary Ware. http://hg.python.org/cpython/rev/f32cb5389a4b New changeset cfd8d99ae645 by Ezio Melotti in branch 'default': #16897: merge with 3.3. http://hg.python.org/cpython/rev/cfd8d99ae645 |
|||
| msg179513 - (view) | Author: Ezio Melotti (ezio.melotti) * |
Date: 2013-01-10 02:34 | |
Fixed, thanks for the patch! While I was at it I also added the other example from the doc. |
|||
| History | |||
|---|---|---|---|
| Date | User | Action | Args |
| 2013-01-10 02:34:40 | ezio.melotti | set | status: open -> closed resolution: fixed messages: + msg179513 |
| 2013-01-10 02:33:36 | python-dev | set | nosy:
+ python-dev messages: + msg179512 |
| 2013-01-09 15:33:43 | zach.ware | set | files:
+ test_bisect_fix.v2.diff messages: + msg179454 |
| 2013-01-09 07:03:15 | ezio.melotti | set | messages: + msg179417 |
| 2013-01-09 06:47:22 | zach.ware | set | messages: + msg179416 |
| 2013-01-09 06:17:01 | ezio.melotti | set | assignee: ezio.melotti messages: + msg179413 stage: resolved |
| 2013-01-08 21:39:05 | zach.ware | create | |