[openstack-dev] [horizon][neutron] tools/tox_install changes - breakage with constraints

Paul Belanger pabelanger at redhat.com
Wed Mar 14 22:41:31 UTC 2018

On Wed, Mar 14, 2018 at 10:25:49PM +0000, Chris Dent wrote:
> On Thu, 15 Mar 2018, Akihiro Motoki wrote:
> > (1) it makes difficult to run tests in local environment
> > We have only released version of neutron/horizon on PyPI. It means
> > PyPI version (i.e. queens) is installed when we run tox in our local
> > development. Most neutron stadium projects and horizon plugins depends
> > on the latest master. Test run in local environment will be broken. We
> > need to install the latest neutron/horizon manually. This confuses
> > most developers. We need to ensure that tox can run successfully in a
> > same manner in our CI and local environments.
> Assuming that ^ is actually the case then:
> This sounds like a really critical issue. We need to be really
> careful about automating the human out of the equation to the point
> where people are submitting broken code just so they can get a good
> test run. That's not great if we'd like to encourage various forms
> of TDD and the like and we also happen to have a limited supply of
> CI resources.
> (Which is not to say that tox-siblings isn't an awesome feature. I
> hadn't really known about it until today and it's a great thing.)
If ansible is our interface for developers to use, it shouldn't be difficult to
reproduce the environments locally to get master. This does mean changing the
developer workflow to use ansible, which I can understand might not be what
people want to do.

The main reason for removing install_tox.sh, is to remove zuul-cloner from our
DIB images, as zuulv3 no longer includes this command. Even running that
locally, would no longer work against git.o.o.

I agree, we should see how to make the migration for local developer
environments better.


More information about the OpenStack-dev mailing list