I am curious as to how people are approaching source code management, especially in environments where there are several developers and where branches are being used as sandboxes. We are currently using RevisionManager, but we haven't been using this in conjunction with branches, has anyone been doing this, and if so, what has your experience with this been? (longer text of our current practice and where we would like to go is at the end.) Is there a standard accepted practice for using source code control with zope (especially for stuff that would be stored in the ZODB), or is Zope most often being used in an environment where this isn't necessary? Comments/ suggestions are appreciated. Sincerely, Kelly McDonald. For historical reasons, almost all of our code/business logic is contained in the ZODB in the form of python scripts/ DTML, etc. (data is stored in a SQL server, postgres) We realize that it would be better to get that stuff out of the ZODB, but we didn't know that when we started, so now we are stuck with a codebase that is all in the ZODB. We have been using Revision Manager to keep our changes in CVS, but we haven't been using branches at all. This has worked OK with two developers, however as we are adding a third and possibly a fourth (sorry, already on staff, not hiring right now!), we see branches as a requirement. I realize that RM has some options for using branches, but it doesn't seem particularly efficient when compared to using cvs/svn directly on the FS. We are currently looking at APE with a file system volume (see my previous post) but are uncertain if APE has been proven yet, even though it has a 1.0 version. For us, the holy grail would be able to have all of our objects on the file-system for source control, and be able to use GUI tools for cvs/svn (no flames please, I prefer CLI, but we have developers who are used to VB,etc) and not take a performance hit for doing any of this. -- Kelly McDonald <w4kpm@adelphia.net>