Convert SVN repos to Git and upload to GitHub

svn2git_improved.sh

#! /bin/bash

# Determine target directory.
TARGET_DIR=$1

# Determine if this is the final processing, i.e. create all Git branches,
# tags, and garbage-collect.
PROCESS_MODE=${2:-FETCH_ONLY}

# Determine default domain for unmapped authors.
EMAIL_DOMAIN=${3:-example.com}

# Set authors file location.
AUTHORS_FILE=$4

# Define current directory as location for authors program.
AUTHORS_PROGRAM_DEFAULT="$(pwd)/authors-prog.sh"

# Set authors program for any unmapped authors.
AUTHORS_PROGRAM=${5:-$AUTHORS_PROGRAM_DEFAULT}

# Optionally, set a starting revision to reduce the size of the Git repository.
STARTING_REVISION=${6:-0}

# Optionally, disable deleting the temporary directory for uploading to GitHub.
DELETE_TEMP_DIRECTORY=${7:-TRUE}

# Determine log filename.
SCRIPT_START=$(date +%s)
echo "Process Mode: $PROCESS_MODE"
echo "Target Directory: $TARGET_DIR"
echo "Authors Program: $AUTHORS_PROGRAM"
echo "Email Domain: $EMAIL_DOMAIN"
echo "Authors File: $AUTHORS_FILE"
echo "Script start (UNIX): $SCRIPT_START ($(date))"

# Switch to target directory.
cd $TARGET_DIR

# Kick off import.
if [ -f "$AUTHORS_FILE" ]
then
  export GIT_EMAIL_DOMAIN=$EMAIL_DOMAIN
  git svn fetch \
  --authors-file=$AUTHORS_FILE \
  --authors-prog=$AUTHORS_PROGRAM \
  --revision=$STARTING_REVISION:HEAD
else
  git svn fetch \
  --revision=$STARTING_REVISION:HEAD
fi

# If this isn't the first time this script runs, fast-forward the actual commits.
git svn rebase

prepare_for_github() {
  local GIT_DIR=$1
  local PREVIOUS_DIR=$(pwd)
  echo "Switching into $1"
  cd $1

  # Create a true Git branch for every "ref" that's not a tag.
  # Exclude: master branch, trunk, branches starting with "trunk", branches containing special character "%"
  git for-each-ref --format="%(refname:short)" refs/remotes/origin | sed 's#origin/##' | grep -v '^tags' | grep -v '^trunk' | grep -v 'master' | grep -v '%' |
  while read BRANCH_NAME
  do
    echo "Create branch name: $BRANCH_NAME"
    git branch $BRANCH_NAME origin/$BRANCH_NAME
  done

  # Create true Git tags for all tags.
  # Exclude: the "origin/tags/tags" tag
  git for-each-ref --format="%(refname:short)" refs/remotes/origin/tags/ | grep -v 'origin/tags/tags' |
  while read tag
  do
      GIT_COMMITTER_DATE="$(git log -1 --pretty=format:"%ad" "$tag")"
      GIT_COMMITTER_EMAIL="$(git log -1 --pretty=format:"%ce" "$tag")"
      GIT_COMMITTER_NAME="$(git log -1 --pretty=format:"%cn" "$tag")"
      git tag -a -m "$(git for-each-ref --format="%(contents)" "$tag")" "$(echo "$tag" | sed 's#svn/tags/##')" "$tag" || exit 1
  done
  
  # Delete all local branches that have already been merged into master,
  # i.e. are neither ahead or behind master.
  git branch --merged master | grep -v 'master$' |
  while read merged_branch_name
  do
    git branch -d $merged_branch_name
  done

  # Garbage-collect
  git svn gc
  git gc --aggressiv

  # Switch back to previous directory.
  cd $PREVIOUS_DIR
}

if [ "$PROCESS_MODE" == "PREPARE_CURRENT_FOR_GITHUB" ]
then
  prepare_for_github $TARGET_DIR
elif [ "$PROCESS_MODE" == "CREATE_TEMP_AND_PUSH_TO_GITHUB" ]
then
  # Create temporary directory and process files in directory.
  REPO_TEMP=$(mktemp -d)

  # Delete temporary files when program exits.
  if [ "$DELETE_TEMP_DIRECTORY" == "TRUE" ]
  then
    echo "Temporary directory $REPO_TEMP will be deleted."
    trap "exit 1" HUP INT PIPE QUIT TERM
    trap "rm -rf $REPO_TEMP" EXIT
  else
    echo "Temporary directory $REPO_TEMP will not be deleted."
  fi 

  # Rsync all files into separate directory
  rsync -a --stats ${TARGET_DIR}/ ${REPO_TEMP}

  # Create all branches in temporary directory.
  prepare_for_github ${REPO_TEMP}

  # Upload everything from temporary directory if
  # 1. Key agent has at least 1 identity and is running and
  # 2. Git remote is added to the repo
  if ssh-add -l && git remote show origin
  then
    git -C $REPO_TEMP push --all --force
    git -C $REPO_TEMP push --tags --force
  else
    echo "Either remote is not set up successfully or SSH key is not working."
  fi
fi

# Log end date.
SCRIPT_END=$(date +%s)
echo "Script end (UNIX): $SCRIPT_END ($(date))"
SCRIPT_DURATION=$((($SCRIPT_END-$SCRIPT_START)/60))
echo "Migration took $SCRIPT_DURATION minutes."