On Tue, Nov 17, 2015 at 2:48 PM, John Chilton <jmchilton(a)gmail.com> wrote:
Thanks for the bug report. Somehow Galaxy isn't installing the
development wheels into the transient Galaxy's virtualenv, I've wiped
out my planemo caches and I can't reproduce this locally.
Perhaps unrelated, but I was seeing something similar
This was despite pip having clearly installed the sqlalchemy-migrate
wheel somewhere (presumably not in the right place though).
(That was using wheels under the Galaxy dev branch, the same
TravisCI setup was working with eggs with the Galaxy master branch)
That was using my own virtualenv on top of the system Python.
I "fixed" this by adopting the TravisCI provided Python virtualenv
(which I'd been meaning to do anyway - my different TravisCI
setups had diverged a bit too much),