i've started my junkcode repository, storing it in CVS. this is less than ideal, as I can see me having work in three locations -- my laptop, my work machine and my home machine. the cvs repository lives on my house server, which should be accessable via internet at any time which should aleviate most of the issues. as i have full access to the canonical respository I can clean it up if things go astray too (something you can't do with sourceforge for example). what i would really like is the ability to push and pull between trees without having to learn arch. I'm not the only one, and Martin Pool has just started work on it.
In the mean time, if I'm getting into issues where I know I'll be with the laptop and want to check stuff into the repository with no net, I'll probably set up something that replicates the junkcode cvs root and modifies
CVS/Root
and uses rsync or something