On Thu, Mar 5, 2009 at 10:13 PM, Robert Turner <r...@flexadata.com> wrote: > Hi Jens, > > You need to branch the codebase. A typical SVN repository is setup as: > > <repos>/trunk > <repos>/tags > <repos>/branches > > Most people will checkout <repos>/trunk and just work off that. What you > need to do is copy the trunk (svn cp/copy <repos>/trunk > <repos>/branches/branch-<version>). Then keep two checkouts: <repos>/trunk > (current versoin) and <repos>/branches/branch-<version> (old version).
You appear to misunderstand the purposes of branches and tags. Branches are designed in order that you can work on something without disturbing the integrity of the main tree, tags are meant to 'tag' a release, i.e. something that was considered worth releasing or going back to. Most people create a branch for each new feature. When they're done, they merge the trunk back into the branch (to pull changes other people have made) and commit the branch code into the trunk. At that point you can kill the branch because it no longer serves a purpose. When the feature is done, if it warrants a tagged release, then tag it. If it doesn't, don't bother. If you need to later work on it, *then* you create a branch from the tag before merging and committing as per normal. As an aside, I find it easier to just check out the repo again into another directory, at a particular revision in the past. Nowadays I use git-svn and just create a topic branch for each change, merging them straight back into the trunk if they're small. There are a few annoying bugs in git-svn that I've been meaning to patch, however. --James ******************************************************************* List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: memberh...@webstandardsgroup.org *******************************************************************