yuppie wrote:
Tres Seaver wrote:
Test failures -------------
Subject: FAILED (errors=1) : Zope-trunk Python-2.4.3 : Linux From: Zope Unit Tests Date: Mon Oct 23 21:35:18 EDT 2006 URL: http://mail.zope.org/pipermail/zope-tests/2006-October/006371.html First, thanks for resuming this collected report. I would much rather have links directly to the broken test run than to the buildbot page.
Second, I can reproduce the failure on my machine, at least after an 'svn up' this morning, but have no obvious clue why it would be broken: the code for local site stuff in Five hasn't changed since May.
The failure showed up after this checkin: http://svn.zope.org/?rev=70383&view=rev
But I didn't have time to track it down, so I opened this collector issue: http://www.zope.org/Collectors/Zope/2198
OMG, why are we using the 3.3 branch HEAD on the 2.10 branch??? This is how Stephan's fix got pulled into Zope 2 automatically and *automatically broke* Zope 2. It also broke Five trunk for some reason (the formlib stuff). Obviously, this is bad. We should stick to specific revisions or tags when pointing to externals. Pointing to HEAD of a branch or trunk is insane as it's an uncontrollable risk. Philipp