[Zope-dev] zope-tests - FAILED: 8, OK: 39

Christophe Combelles ccomb at free.fr
Tue Nov 15 02:52:28 UTC 2011


Le 14/11/2011 23:24, Tres Seaver a écrit :
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 11/14/2011 04:50 PM, Christophe Combelles wrote:
>> Le 14/11/2011 01:52, Christophe Combelles a écrit :
>>> Le 13/11/2011 14:22, Tres Seaver a écrit :
>
>>>> If somebody who cares about the KGS could look at these failures
>>>> and figure out why knew versions are slipping into the mix, that
>>>> would be great.  Otherwise, we should just quit testing it.
>>>
>>> The initial build process of the kgs has no versions pinned, so it
>>> was being built with some of the latests libs that fail under
>>> python2.4.  I have added versions in the toplevel buildout.cfg so
>>> that it shouldn't fail again.
>>>
>>> I don't know if someone is interested but we have enough minor
>>> versions upgrade available to be able to release a maintenance
>>> versions of the KGS: 3.4.2
>>
>> fixed
>>
>> http://buildbot.afpy.org/kgs3.4/builders/Python2.4.6%2064bit%20linux/builds/441
>
> I
>>
> don't understand why r123334 was needed to "fix" the KGS buildout:  I
> thought the KGS was "pinned" by virtue of using its 'versions.cfg' file:
>
>   http://download.zope.org/zope3.4/3.4.1/versions.cfg

The problem doesn't come from the KGS itself, but from the package that creates 
the KGS, I mean zope.release. (which itself depends on the logic of zope.kgs).
During the release process, zope.release generates a buildout that will be used 
to launch tests. So there are 2 buildouts : the one from zope.release itself, 
and the one generated by the generate-buildout command.


>
> (Actually, I don't know what the intended relationship is between that
> file and http://download.zope.org/zope3.4/3.4.1/controlled-packages.cfg ).

The versions.cfg is generated from the controlled-packages file, and contains 
only the latest versions.  controlled-packages was supposed to track all the 
compatible versions (even older ones) of the KGS.
But it happens to be mathematically wrong and impossible, because the cobinatory 
explosion of versions makes it unrealistic to handle all the combinations.


>
> Actually, I thought we were supposed to have a PyPI-like index which
> ensured that no unwanted versions would creep in, e.g.::
>
>   http://download.zope.org/zope3.4/3.4.1/index/
>
> Can somebody who remembers how that dance was supposed to work enlighten us?


Everything is here
http://svn.zope.org/zope.release/branches/3.4/README.txt?rev=89181&view=auto

This is much simpler with the ZTK now.

>
>
> Tres.
> - --
> ===================================================================
> Tres Seaver          +1 540-429-0999          tseaver at palladion.com
> Palladion Software   "Excellence by Design"    http://palladion.com
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.10 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
>
> iEYEARECAAYFAk7BlRMACgkQ+gerLs4ltQ7YhwCg2g8n/s26u0z5FvkrgwtTV+z8
> ExQAoLrqxIAdMhQ+G1skUR2o7qa6LoOA
> =BrQZ
> -----END PGP SIGNATURE-----
>
> _______________________________________________
> Zope-Dev maillist  -  Zope-Dev at zope.org
> https://mail.zope.org/mailman/listinfo/zope-dev
> **  No cross posts or HTML encoding!  **
> (Related lists -
>   https://mail.zope.org/mailman/listinfo/zope-announce
>   https://mail.zope.org/mailman/listinfo/zope )


More information about the Zope-Dev mailing list