Thanks for the comment Ben.
Ben Prew saidWe have a similar situation where I work in that we have several disparate sites, but we want to use the same layout. We solve this by maintaining a symlink that we check in to the repository that points to the shared directory, so when we have to update parts of the shared code, we can update them, and the disparate sites get the updates because they just contain a symlink to the shared location, as opposed to the code itself.
When you mentioned not having shared files, I realized it was because you're using Windows, which doesn't let you create symlinks and check those into svn, but its a nice way to solve that problem.
Do the symlinks work at the file or the directory level? If at the directory level you can get a fairly similar result using svn:external.
The nice about the SourceSafe feature was that if the share was active SourceSafe was in effect telling you that the file was meant to be the same across a number of sites. Sometimes we have to fork just one file for one web site and SourceSafe would preserve knowledge of that fork in the repository.
In Subversion we will only know that a given file is different on a web site. Knowledge about which web sites are meant to have identical versions and which are meant to have different versions must be held somewhere else so is far more likely to get lost, or be ignored.
I can see that some changes needed for some web sites stand a chance of getting overwritten and some changes won't be propogated properly. More complex procedures around file updates will go some way towards mitigating the problems, but there is a fairly simple technical fix — at least from our point of view as Subversion users, even though I'm sure the Subversion code to handle it will be anything but simple.
