My current SuperCollider development workflow is based on this setup

  • I use sourcetree (I’m on OS X)
  • I forked SuperCollider to my github account (atm. http://github.com/lfsaw/supercollider)
  • I renamed the remote pointing to the official SuperCollider repo (http://github.com/supercollider/supercollider) upstream
  • I added my personal fork at http://github.com/lfsaw/supercollider as origin

If developing SuperCollider, I

  • check out recent upstream/master (atm. upstream/3.9 because we’re in release mode) and compile it,
  • branch a topic/branchname and commit my changes to it,
  • push the branch to origin,
  • create a pull request for the branch and wait for reactions.

If working on my own stuff (“composing”), I

  • switch to master
  • recompile SuperCollider (i.e. cmake), though most of the times this is not needed since many changes in the language are happening in .sc and .schelp files.