I agree that having testing done is important, but just deploying a char and having it successful and a charm actually working won't be the case for most of the charms because a lot of the current charms do its setup in hooks other than install, especially with charms that require relations. So even as a short term solution I'm not sure how effective it would be. What about having a test folder in each charm and having authors write out tests in bash or another such language that would go through the process of getting req charms, deploying them, adding relations, changing configs, then gathering all the state information from status and comparing outputs with expected outcomes. Or would that be too much/out of scope for charms? <div style="font-family: arial, sans-serif; font-size: 12px;color: #999999;">-- Marco Ceppi Sent while mobile</div><hr align="left" style="width:75%">On Nov 30, 2011 2:19 PM, Mark Mims <mark.mims@canonical.com> wrote: Hey, so testing charms correctly is hard. What about a partial solution to automated charm testing that can be delivered sooner? There's actually quite a bit of value in just insuring that install hooks complete. I think it's worth dedicating a system to just do regular deploys, wait for a corresponding 'started' state, and report on the outcome... no stacks, no relations. We'll obviously need to build on this over time, but just this one step would vastly improve the maintainability of the charm store. Thoughts? -- Mark Mims, Ph.D. Canonical Ltd. mark.mims@canonical.com +1(512)981-6467 -- Juju mailing list Juju@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju