dotfiles/.local/bin/dataindex

#!/bin/sh

usage () {
	cat <<EOF>&2
Usage: ${0##*/} [OPTIONS] FOLDERS

Output index of folder hierarchies to stdout. This is used as a small "backup"
purpose. In case of data loss, it is still possible to get the full file list
thanks to the indexes.

Version control and encfs-encrypted folders are skipped.

Options:

 -f: When outputting to a file, overwrite if it exists.
 -w: Output each index to individual files in current folder. It uses the folder
     name as basename.
EOF
}

opt_noclobber=true
opt_file=false
while getopts ":fhw" opt; do
	case $opt in
	f)
		opt_noclobber=false ;;
	h)
		usage
		exit ;;
	w)
		opt_file=true ;;
	\?)
		usage
		exit 1 ;;
	esac
done

shift $((OPTIND - 1))

if [ $# -eq 0 ]; then
	usage
	exit 1
fi

## 'realpath' is required in case argument is ending with '.' or '..'.
if ! command -v realpath >/dev/null 2>&1; then
	echo >&2 "'realpath' not found"
	exit 1
fi

for i ; do
	[ ! -d "$i" ] && continue
	[ -e "$i"/.encfs*.xml ] && continue

	## We strip "./" from find's listing since we don't need it. We could avoid
	## printing it in the first place, but there are several shortcomings:
	## - Find over '.*' and '*', is bad practice since if will fail on
	## non-existing files or for files beginning with a dash.
	## - The 'printf' command in find is for GNU find only.

	## 'LC_ALL=C sort' is required to make sure to output is consistent across
	## different systems.

	echo >&2 "$i"
	## The two following lines do the same for the same time cost. The former is shorter.
	# find "$i" -type f | awk -v str="$i" '{l=length(str)+2; print substr($0, l)}' | LC_ALL=C sort > "$OUTPUT"
	(cd -- "$i" && find . \( -path '*.git*' -o -path '*.svn*' -o -path '*.hg*' \) -prune -o -type f) | sed 's/^.\///' | LC_ALL=C sort | \
		if $opt_file; then
			i="$(realpath -- "$i")"
			base="${i##*/}"
			output="$base.index"
			[ -e "$output" ] && $opt_noclobber && output="$base-$(date +%F-%T).index"
			echo "$output"
			cat > "$output"
		else
			cat
		fi
done