How long before we end up with 2.2.4 a3 and 2.2.4 b1 with no-one really knowing what's going on? Taking what is effectively another point onto the version number doesn't really help. I can only think of 2.1.7 that was withdrawn, and as long as 2.2.4 comes along pretty swiftly, there shouldn't really be a problem, should there?
Well, some people perceive it to be problem (and I can't really blame them). What we _really_ need is some regression testing infrastructure. We have started taking some baby steps in that direction with the inclusion of PyUnit, but it will take some time and resources to get a meaningful amount of tests. Hopefully the recent Good Things happening here at DC will mean that the necessary resources will become available to make that happen. In the meantime, since Brian is a poor regression test rig, I plan to have a (generally very short - like a week) beta period for third-dot releases. One beta should usually do it, and it will allow at least for more coverage than I can easily contrive from my desktop as a sanity check. When the test infrastructure is more mature we could revisit this though. Brian Lloyd brian@digicool.com Software Engineer 540.371.6909 Digital Creations http://www.digicool.com