mirror.sh

#!/bin/sh

set -eufo pipefail

if [ "$#" -ne 2 ]; then
    echo "usage: $0 source_repo_url target_repo_url" >&2
    exit 1
fi

SOURCE_URL="$1"
TARGET_URL="$2"
WORKDIR="$(mktemp -d)"

echo "Cloning from ${SOURCE_URL} into ${WORKDIR}..."

git init --bare "${WORKDIR}"
cd "${WORKDIR}"

git config remote.origin.url "${SOURCE_URL}"
git config --add remote.origin.fetch '+refs/heads/*:refs/heads/*'
git config --add remote.origin.fetch '+refs/tags/*:refs/tags/*'
git config --add remote.origin.fetch '+refs/notes/*:refs/notes/*'
git config remote.origin.mirror true
git fetch --all

echo ""
echo "Cloned to ${WORKDIR}; pushing to ${TARGET_URL}"

git push --mirror "${TARGET_URL}"

echo ""
echo "Cleaning up temporary directory ${WORKDIR}..."

rm -rf "${WORKDIR}"

echo "Done."

@maxenc7 I wrote this 5 years ago. I don't remember precisely what my goals were but I suspect that it was intentionally done to limit the refs that were mirrored. If you git clone --mirror a repository on GitHub, you'll get all the pull request refs (refs/pull/<number>/head and refs/pull/<number>/merge). And all the objects that are reachable from those pull request refs. So specifying only a ref refspecs will give you the state of the repository (branches, tags, notes) without any of the things that were never actually merged into a branch.

Why I wanted to do this, I don't know, and maybe there's a way to make git clone --mirror do this as well, I don't know that either. 🤷

Indeed - I wrote about it here: https://www.edwardthomson.com/blog/advent_day_23_removing_large_binaries_with_lfs.html but didn't actually link to the script. And somebody else put some helpful comments in a fork of this gist so that I would know what I was thinking. https://gist.github.com/Fishezzz/12e94d9a06daaa9af461c897583fe3fd

(Thanks @Fishezzz)