As part of the switchover to git from svn, I'd like to clean large, no longer interesting files from the history (eg large data file that were added and then removed). Removing those files from the history reduces the size of the repository by about 50%. Since this involves rewriting the history, it would require some action by people who have an existing copy of the git repository. If you do and have no changes of your own in it, you can - simply clone a new copy - do "git pull -f origin develop" to pull a new copy of the develop branch (similar for master) If you have branches in your clone, then the easiest way is to rebase you changes against the updated repository. I can provide instructions for this, if desired.
The main effects of shrinking the repository are: - its size on disk will be reduced from ~1 gig to ~.5 gig - clone will take half as long Otherwise git is pretty good at being fast no matter what.
Does this seem desirable to people? --Daniel