This is accademic because I already did what I describe below, but I wonder if there would have been a better solution.
I had this long branch, which I'll call xp15 (a Windows XP branch amid NT branches actually), which was production-level code. On the trunk, I had a lot of experimental code which was supposed to go production but never did. Having learned since creating that arrangement the wisdom of doing things the other way (branch the experiments, leave the trunk a production source), I wanted to make HEAD again be the main development--in this case, continue from the tip of the xp15 branch; and I wanted to move all the experimental stuff off onto its own branch, which I'll call exp here. Sequence of events as best I can manage from memory: Trunk development up to initial branches. Split off xp15 branch (and others). Commit to trunk with experiments that never made it out. Commit to xp15 (and others) and release from there. ...a lot of time passes during above two items ... <<begin discontinuity>> Check out a copy of the (experimental code) trunk. cvs tag exp_base cvs tag -b exp cvs update -A (to make sure of no sticky tags/branches in sandbox) Check out the tip of the xp15 branch in a separate sandbox. Make the trunk sandbox a current copy of the xp15 sandbox. Send a huge commit consisting of all adds, removes, and mods to make trunk become equal to tip of xp15. <<end discontinuity>> Continue working on trunk with XP-related commits. Continue working on exp branch with experimental stuff (maybe). This worked pretty well but of course made the process of tracing XP development history a little interesting, and the same for the experimental code history if it resumes in the future. I know no better way to handle this, but I thought I'd see if someone else does. -- Doug Lee [EMAIL PROTECTED] SSB + BART Group [EMAIL PROTECTED] http://www.bartsite.com "Characters live to be noticed. People with character notice how they live." -- Nancy Moser _______________________________________________ Info-cvs mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/info-cvs
