Archive Legacy Github Repos with Subtree

If you're like me your github account is littered with hundreds of legacy experimental repos that you don't want to toss.

git subtree is the magic needed to consolidate them into an archive, preserving all of the commit history . Once consolidated, all of those legacy repos can be archived to your cloud storage and deleted from github.

Clone the Archive Repo

First step is to choose (or create) your archive repo that will contain all of your experiments

USERNAME=tony

REPO=archive

git clone [email protected]:$USERNAME/$REPO.git

cd $REPO

Clone one or more Experiment Repos (you would like to delete)

USERNAME=tony

REPO=experiment-web

git clone [email protected]:$USERNAME/$REPO.git

Add the local experiment-web as a "remote"

note: Git remotes can be copies from a local directory. This is the magical source of subtree's power to fan-in or fan-out repos

cd archive

name=experiment-web ; git remote add subtree/$name ../$name  

Do this any number of times for all of your experiment Repos

Subtree The Repo and Test the Results

# stay within Archive

name=experiment-web

git subtree add -P $name subtree/$name master # or any branch

Test that all the contents match exactly

cd $name

diff -r . ../../$name

# Only in ../../experiment-web: .git

Test that the subtree'd log matches the log of the original repo

git log --graph -4

*   a19d970 - (HEAD -> master, tonymet/master) Add 'oh-my-zsh/' from commit '9f3f282699dccc11221debbdbec4be654c63ac83' (11 minutes ago) <tony>

|\

| * 9f3f282 - (subtree/oh-my-zsh-origin/anthony) macbook (14 minutes ago) 

Archive and Delete The Repo

Now you can archive the repo to your cloud storage ( Google Cloud , etc).

# Save to TGZ file and copy to your Google Drive

name=experiment-web && tar -zcf archive/$name-`date +%Y-%m-%d`.tgz $name &&  rm -rf $name  

# Visit repo on the web to delete