I never put a paper up on github until after it has been published, and I've seen considerable push-back from fellow scientists at the suggestion we consider it.

Despite the pervasive influence of github, it is quite easy to spawn local git/svn/hg repos.