I'm using CVS almost exclusively for web development and am trying to determine how I should handle branching.
My current method uses the typical dev/stage/production areas. Each developer has his own sandbox. There is a single "stage" sandbox and the live site is simply a checked out copy of the repository. The usual chain of events is... 1. Initial development work is completed and the staging site is updated for review by client. 2. Once approved, a copy is checked out as the "production" site. 3. I create a "release-1_0" tag on the trunk and development continues. What I'd like to do is determine the best way to handle development, branching and tagging from this point forward. My first thought is to create a "release-1_0-updates" branch off the "release-1_0" branch and move the stage and live sites onto that branch. Developers then can either continue working on the main branch heading toward the next major update to the site, or move their sandbox to the updates branch for minor changes and bug fixes. When release-2_0 is ready, simply merge the release-1_0-updates and main branches, tag and branch it as "release-2_0-updates", rinse and repeat. It seems similar to the normal software development cycle, with a few differences such as files (content) that get modified on the live site by the client (it can happen), which is why the live site is actually a working directory. This sound reasonable? -- Jack Baty Fusionary Media - http://www.fusionary.com/ _______________________________________________ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs