Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Sanity check for setuptools 15.0 #242
I've tagged and uploaded a b1 release of setuptools 15.0. This release re-attempts a change that once failed due to assumptions in buildout which I believe have since been corrected.
Can someone test and confirm that the 15.0b1 release is suitable and won't likely create unmanageable breakage in the buildout community?
Thanks very much for checking in.
I tried it on a few of my buildouts and it seems to work fine, although my buildouts tend to be pretty tame. :)
I tested this by copying the release to my download cache (the
Hopefully, some others here will try out this release too.
I tried a few buildouts and it went fine.
Let's try to break it. Ah, setuptools 15.0b1 in combination with zc.buildout 2.2.1 will fail. zc.buildout 2.2.2 or higher works fine. Here is a buildout config to test it:
It goes wrong like this:
Reason it works fine with setuptools 14.3.1 is because in a pdb you get this:
And with the setuptools beta you get this:
(Note that there is nothing wrong with the requirement here. The exception simply means that zope.testrunner is not found locally, so zc.buildout enters a code section where it tries to get it from PyPI.)
So there are certainly versions of zc.buildout that will break with this new setuptools version. At least 2.2.0 and 2.2.1, maybe others. But newer versions in the 2.2.x line work fine. And I see that the 2.1.x series wants distribute instead of setuptools, so that series seems no concern.