From: Carsten Ziegeler
Im really tired of discussing this (it's not your fault, Reinhard).
I suggested again and again to start a new repository for 2.2 (which is equivalent to starting a 2.2 branch - or a 2.1.x branch), but I always got the reply "let's see if we really need 2.2, let's stick to 2.1 for now." Now, only 5 days later people ask for it. So, if someone feels that our current approach is wrong, start a new thread; I will not do it. BTW, each release is tagged and we can always create a maintainance branch afterwards - but that wouldn't be the way we decided to use the cvs months ago...
IMO we should be able to release 2.1.x version in the future. Mabe I'm a bit afraid of the changes (blocks, fortress, virtual sitemap components) and how long it takes us to release stable versions of Cocoon (not beta, not milestone or something else) again if we don't branch or fork our repository.
I think so many people (of course including me) were waiting for 2.1 and I really want to use it for some time in production without doing alpha/beta testing of 2.2 and I like to have "support" of bugfix releases of 2.1 (see the security holes Sylvain fixed recently, maybe we can release a stable Cocoon Forms, ...).
Am I (and Carsten) alone with this opinions?
No, I agree too. I think the move to Fortress makes a 2.2 repo wise. Berin is optimistic about the chances of mitigating problems with the transition, and I trust in that. But it is IMHO a significant enough architectural change which can/will have un-forseen consequences across the user-base.
If we don't, we risk putting people in a Microsoft-like situation: in order to upgrade to fix bug A they need to accept bug B introduced in the new version.
Exactly how that's achieved is less important. Of the less than optimal options, I think a 2.2 repo is the best bet.
Geoff
