Sorry Jim, I'm working on both Radiant and Refinery projects at the moment, and getting a little confused by the "R" words. I meant "does radiant come with cucumber scenarios", but I think you got that.
It's kind of strange testing my Radiant app. When I run rake spec, it runs *everything*. Heaps of specs from Radiant itself, for models, controllers, helpers, etc. It runs specs for all of the extensions. Interestingly, the entire process fails if any of the radiant specs fail, but spec failures in the extension suites don't cause the "rake spec" process to return failure as a whole. The whole suite takes ages, over five minutes.
When I add extensions, they inevitably monkey patch Radiant, and then it's really hard to keep the specs passing. Extensions that do things like reorder pages inevitably cause radiant specs to fail (since it expects a different order). This is to be expected I imagine. It's sort of odd to run the tests for Radiant and all the extensions on one big app. How can they all be expected to pass? Since I'm now to the Radiant app I'm working with, I've carefully kept them all passing, and it took days. I've now got all the extensions I need installed, and a couple of the extensions have failing specs but Radiant doesn't, and so the process as a whole passes.
More sensible would be if only my application itself ran specs using the app code. Radiant specs should run without my app, or extensions loaded. Extension specs should run with radiant loaded, but not anything else. I shouldn't typically run Radiant or extension specs when I'm developing my app, only if I'm working on radiant itself, or an extension. I should be able to test extensions from within their own repositories without having to put them into an app.
Refinery (and I do mean refinery this time) does this by using the rather crude approach of putting a dummy rails app into the spec directory of Refinery itself. So, it doesn't copy that into the gem, but it is used for running the specs. I'm a bit less sure how extension specs work.
It seems to me like poor testing processes are a struggle for Radiant. It's a more feature rich CMS than Radiant, with more useful extensions available. However, the extensions typically get orphaned, or change without a proper upgrade path. My experience of upgrading a refinery 0.8 app, with custom extensions, to a 1.0.1 app has led me through a huge number of github forks of extensions, and shown some pretty hideous code in some of the extensions. I've seen work by two separate consultancies doing Radiant work professionally and yet not writing tests, perhaps because it seems too hard.
Anyway, pretty easy to pick on things from the outside. Obviously a huge amount of work has gone into things, and I don't want to be too critical. I hope the above thoughts are of some constructive use to folks.