I have recently been asked to revive an old project. Way back when I used to use bk for tracking changes. But today, I don’t even have a working bk tree.

Moving the history to git is easiest done by taking the tarballs I’ve published and creating a commit per tarball.

Below is a simple script that will do just that.

To run it you’d do something like:

    mkdir my-repo
    cd my-repo
    git init
    ../import ../*.tar.gz

And here is the import script:

    #!/bin/zsh
    set -e

    commit=
    for tar in $@ ; do

            echo >&2
            echo >&2 "# Importing: $tar"
            [[ -f $tar ]] || exit 1
            
            base=$(basename $tar)
            date=$(stat --format="%y" $tar)
            
            git rm -qrf . || true
            tar -xz --strip-components=1 -f $tar
            git add .
            
            tree=$(git write-tree)
            commit=$(GIT_AUTHOR_DATE="$date" git commit-tree $tree ${commit:+-p} $commit <<<"import of $base")
            
            git reset --hard $commit
    done    

    echo >&2
    echo >&2 "# SUCCESS"

_

NOTE: the tar command will remove the leading path with the assumption that your project tar balls contain a single leading directory for all the files… as it is traditionally done.

WARNING: you don’t want to run this script from a repo that already has commits. This script forces the first tarball passed as the argument to become the first commit in history. The current branch will contain no history other than what is in the tarballs.