[Distutils] Failing test on zc.buildout.

Gary Poster gary.poster at canonical.com
Thu Jul 22 13:01:21 CEST 2010


Hi Lennart.  Some of this has been discussed on the list before, and I'm afraid you have duplicated some work with the zope.testing changes.  

Specific to this question, the bootstrap on trunk is the wrong bootstrap because of the virtualenv problems and the fact that people use externals to link to bootstrap.

Please see the thread started by Martijn for follow-up, or I'm happy to talk to you on IRC (which might be nice if you can help me dupe the problem you reported against 1.5.0b1, as I wrote you a week or two ago).

In related news, I have decided on a way forward with releasing buildout.

- In my betafix branch, I have already made zc.buildout work with virtualenv as far as I know, as announced before and confirmed by at least one person.
- In case anyone encounters issues with the next beta release, I will create an "aspirin" release: a zc.buildout 1.4.4 that prefers itself if a version is not specified, and a bootstrap that prefers 1.4.4 if it is not specified.  This should give people a way to alleviate their immediate problems if they encounter any: people can just switch to the "aspirin" bootstrap and stay in the 1.4.4 line of zc.buildout until problems are resolved.
- I will make 1.5.0 and its associated bootstrap automatically prefer final releases *for the build system* (zc.buildout and recipes) so that we won't encounter this problem again.

Once I do this, I'll merge my betafix branch back to trunk (fixing tests on trunk) and make a b3 release.

Thanks to Benji York for helping me figure out this plan of attack.

Gary



On Jul 22, 2010, at 5:51 AM, Lennart Regebro wrote:

> I was going to update zc.buildout to use doctest instead of
> zope.testing.doctest and zope.testrunner instead of
> zope.testing.testrunner, and I noticed trunk has some failing tests.
> They all use the --setup-source option, but:
> 
>  bootstrap.py: error: no such option: --setup-source
> 
> So there is something wrong there. I'll probably comment those failing
> tests out for the moment, unless somebody fixes them or tells me to
> remove them completely. :-)
> 
> -- 
> Lennart Regebro: http://regebro.wordpress.com/
> Python 3 Porting: http://python3porting.com/
> +33 661 58 14 64
> _______________________________________________
> Distutils-SIG maillist  -  Distutils-SIG at python.org
> http://mail.python.org/mailman/listinfo/distutils-sig



More information about the Distutils-SIG mailing list