
hi Paul,
our team has just moved from IBM's RTC to Git not long before, with an unfinished project. We downloaded Codebase Origin from the original stream, Codebase A from Stream A, and Codebase B from Stream B. One developer made all patches between Origin and A, and then between Origin and B. A manifest was produced, which devides Codebase Origin into 600+ projects. These projects' codes were pushed to the server, stored in Branch Origin of each project. We then created Branch A from Branch Origin with the patches between Origin and A. The same was done to create Branch B. Manifest A and Manifest B were produced by adding 'revison=Branch A' and 'revison=Branch B' in the projects which had been applied with any of the patches. The manifests themselves are managed by a seperate git with Branch Origin, A and B.
So now if we want to download Codebase A, we just run 'repo init -u someUrlOfManifest -b BranchA' and 'repo sync' to download Codebase A. If some specific project is needed, just run 'git clone someUrlOfProject [-b BranchName]'. You can get detailed info from any git related websites or books.
If you need any further or detailed info pls mail me at [email protected]