Merge lp://staging/~lifeless/zope.testing/subunit into lp://staging/zope.testing

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Robert Collins wrote:
> Robert Collins has proposed merging lp:~lifeless/zope.testing/subunit into lp:zope.testing.
>
> Requested reviews:
> ZTK steering group (ztk-steering-group)
>
>
> Use subunit's extended API more carefully, avoiding passing details
> objects into parameters expecting error tuples.
>
> I don't know if tests pass - the getting started instructions just
> blowup for me, and it didn't look possible to do 'python -m
> unittest.run zope.testing'.

The "standard" way to run the tests would be via buildout:

 $ /opt/Python-2.6.5/bin/python bootstrap.py
 $ /bin/buildout
 $ bin/test --all

I would actually prefer if the tests became / stayed runnable via the
setuptools 'test' command:

 $ /opt/Python-2.6.5/bin/python setup.py test

Tres.
- --
===================================================================
Tres Seaver +1 540-429-0999 <email address hidden>
Palladion Software "Excellence by Design" http://palladion.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkwGx24ACgkQ+gerLs4ltQ4NfgCgvKaifq96pgLATTrX7OD9Admo
FsUAn0P1sNNl/5k/UK5pKbTMN7fftKQR
=PkJm
-----END PGP SIGNATURE-----

Well, I ran python bootstrap.py and it went boom; at that point I
committed and submitted this proposal ;)

The boom was about accessing some setuptools egg, I think.

I fixed the "python bootstrap.py and it went boom" problem by updating the bootstrap.py.

/me adds a line to the why-I-don't-like-copies-of-bootstrap.py-everywhere list.

The fact that this patch doesn't touch any tests suggests that the subunit output is missing some test cases. I'd like to see some tests that fail without the code changes and pass afterward.

I agree there are tests missing; I don't see that holding improvements
of untested code hostage to tests is a good balance.

On Tue, Jul 20, 2010 at 11:15 AM, Robert Collins
<email address hidden> wrote:
> I agree there are tests missing; I don't see that holding improvements
> of untested code hostage to tests is a good balance.

I don't expect the patch to include all the missing tests, but a
regression test would be appropriate. Since you produced the patch, I
assume you can reproduce the behavior that the patch corrects.
--
Benji York

On Tue, Jul 20, 2010 at 5:31 PM, Benji York <email address hidden> wrote:
> On Tue, Jul 20, 2010 at 11:15 AM, Robert Collins
> <email address hidden> wrote:
>> I agree there are tests missing; I don't see that holding improvements
>> of untested code hostage to tests is a good balance.
>
> I don't expect the patch to include all the missing tests, but a
> regression test would be appropriate.  Since you produced the patch, I
> assume you can reproduce the behavior that the patch corrects.

No, I can't- I just know the API in question.

Well, I'm not happy with the lack of tests, but I suppose that the best thing I can do is merge this as is -- which I did in revision 117583.