On Fri, Oct 29, 2010 at 10:01 AM, Martin Pool <[email protected]> wrote: > Looking into this a bit with Deryck, i suspect part of the problem is > that the feature flag code is pretty careful to only read from the > database just once per featurecontroller lifetime (ie per web request) > and updates into the database are not causing it to flush its cache. > If that's correct it should be easy to fix. It's odd that it depends > on unisolated behaviour between tests though. >
I've handed this work back to gmb, since he's available for the work and I have a all my stuff post-UDS to attend to. FWIW, I found that using the context manger for feature flags, rather than the fixture Graham got from Maris, was less prone to this issue. One test passed, but another seemed to have the same issue. So better, but not fixed completely. Cheers, deryck -- Deryck Hodge https://launchpad.net/~deryck http://www.devurandom.org/ _______________________________________________ Mailing list: https://launchpad.net/~launchpad-dev Post to : [email protected] Unsubscribe : https://launchpad.net/~launchpad-dev More help : https://help.launchpad.net/ListHelp
