In my ongoing quest to explore a variety of ways of managing config files, I believe I have found a way that is attractively simple.
Step one: git clone
cd ~
git clone -c status.showUntrackedFiles=no -n --separate-git-dir .git $REPO_URL tmpdir
rm -r tmpdir
Rationale:
-
status.showUntrackedFilesis set to "no" so that futuregit statusrequests only show files that were intentionally tracked withgit addandgit commit -
-nmeans no checkout. We aren't ready for it yet. -
--separate-git-dir .gittakes some explanation. It is impossible togit cloneinto a non-empty directory without some extra steps. This trick does it in one step (two if you count the deletion of the throwaway directory). Tell Git to use a separate Git directory but then, sneaky miscreants that we are, we name the directory the default:.git - The undesirable side effect is an extra directory
tmpdirthat has a single.gitfile of no consequence. The entire directory can safely be removed.
Step two for non-empty repo: git checkout
git checkout
Deal with any file conflicts. For instance, you might backup and remove an existing .bashrc. Then run git checkout again. If you are sure that overwriting is OK, you can pass the force flag:
git checkout -f
Step two if new (empty) repo: add files and push
If this is a brand new setup, your repo is likely empty. Add some files:
git add .bashrc
git commit -m "initial commit of Bash config"
Repeat as necessary with additional files and directories. Then push:
git push
Step three: maintain
Keep your files up to date with git add, git commit, git push. Pull remote changes with git pull. In other words, manage your home directory as you would any other Git repo. Please avoid git add . as that will track every file in your home directory, an undesirable endeavor.
If this spawns other creative ideas or optimizations, feel free to post in the comments!
Top comments (0)