Merges multiple git repositories into a single one

merge-git-repo.sh

#!/bin/bash

# See: https://stackoverflow.com/questions/1683531/how-to-import-existing-git-repository-into-another
# https://help.github.com/articles/about-git-subtree-merges/
# https://github.com/git/git/blob/master/contrib/subtree/git-subtree.txt 
# https://git-scm.com/book/en/v1/Git-Tools-Subtree-Merging
# git filter-branch
# git read-tree
# https://unix.stackexchange.com/questions/280217/how-to-replay-git-repository-history-into-subdirectory


###########################
# SYNOPSIS
###########################
USAGE="./merge-git-repo.sh <target_repo_url> <child_repo_1> [, ...]"
#
# CAVEATS:
# - Assumes an empty target directory
# - Potential path collision in non-master branches
#
# TODO: how to handle source branches without collisions.
# Start with empty destination repo
# For each source repo
#    Clone repo
#    Rewrite entire history with path prefix
#    Commit, but not push
#
# Clone destination repo
#    For each source repo
#       Add source remote into destination repo
#       For each source branch
#          Create local branch (if necessary)
#          Replay source branch history into destination
# 
#    For each source repo branches
#       Create a branch matching 

function merge_child_repo {
	SOURCE_URL="$1"
	SOURCE_BASENAME=`basename "$1" ".git"`
	SOURCE_REMOTE="${SOURCE_BASENAME}-remote"
	SOURCE_BRANCH="${SOURCE_BASENAME}-branch"

	echo -e "\n\t*** Fetching ${SOURCE_URL} as remote ${SOURCE_REMOTE}"
	git remote add ${SOURCE_REMOTE} ${SOURCE_URL}
	git fetch ${SOURCE_REMOTE}
	git checkout -b ${SOURCE_BRANCH} ${SOURCE_REMOTE}/master

	echo -e "\n\t*** Move all git files to subfolder ${SOURCE_BASENAME}/"
	mkdir -p ${SOURCE_BASENAME}/
	FILES=`git ls-tree ${SOURCE_REMOTE}/master --name-only`
	for FILE in $FILES
	do
		echo -ne "\t${FILE}";
		git mv $FILE ${SOURCE_BASENAME}/
	done
	echo -e "\n\t*** Done.\n"
	git commit -m "Moved child repository ${SOURCE_URL} to subdirectory ${SOURCE_BASENAME}"

	echo -e "\n\t*** Merging branch back into master"
	git checkout master                
	git merge ${SOURCE_BRANCH} --allow-unrelated-histories   # should add ZZZ/ to master
	git commit
	git remote rm ${SOURCE_REMOTE}
	git branch -d ${SOURCE_BRANCH}                # to get rid of the extra branch before pushing
	#git push                                    # if you have a remote, that is
}



##########################
# CONFIGURATION
##########################
if (( $# < 2 )); then
    echo $USAGE
    exit 1
fi


##########################
# MAIN
##########################
DEST_URL="$1"


echo -e "\n\t*** Fetching ${DEST_URL} to dest/"
rm -rf dest/
git clone ${DEST_URL} dest/
cd dest/

# Loop through source URLs
shift
for SOURCE_URL in "$@" ; do
  #echo "$SOURCE_URL"
  merge_child_repo ${SOURCE_URL}
done