A bash script to list blobs (hash id and path) in a git repo

git-blob-ids.sh

#!/bin/bash
#
# git-blob-ids: List blobs (hash id and path) in a git repo
#
# Usage: git-blob-ids [-s] [-r REF] [-p PATH]
#
# -r REF an optional reference to a commit (hash or symbolic like HEAD), default: HEAD
# -p PATH an option path to a subtree in repo, default is blank for root tree
# -s shortens blob hash ids to 6 characters
#

function short { echo $@ | awk '{print substr($1,0,6);}'; }
function nochange { echo $@; }

ref=HEAD
path=
idfunc=nochange
while getopts "sr:p:" options; do
  case "${options}" in
    s)
      idfunc=short
      ;;
    r)
      ref=${OPTARG}
      ;;
    p)
      path=${OPTARG}
      ;;
  esac
done

if [[ ${#path} > 0 ]]; then
  tree=$ref:$path
else
  # start from root tree
  tree=(`git cat-file -p $ref | head -1 | awk '{print $2}'`)
  tree="${tree}|"
fi

# tokenize by line in git command output
IFS='
'

# breadth-first iteration of sub-trees
while [[ ${#tree} > 0 ]]; do
  el=`echo $tree | awk '{print $1}'`
  t=`echo $el | awk -F"|" '{print $1}'`
  path=`echo $el | awk -F"|" '{print $2}'`
  if [[ ${#path} > 0 ]]; then
    pfx=$path/
  fi
  # pop
  tree=`echo $tree | sed 's/[^ ]* *//'`
  for l in `git cat-file -p $t`; do
    if [[ $l =~ "blob" ]]; then
      blob=`echo $l | sed 's/.* blob //;s/\t.*//'`
      blob=`$idfunc $blob`
      lpath=`echo $l | sed 's/.* blob //;s/.*\t//'`
      # output blob info
      echo $blob $pfx$lpath
    elif [[ $l =~ "tree" ]]; then
      id=`echo $l | sed 's/.* tree //;s/\t.*//'`
      lpath=`echo $l | sed 's/.* tree //;s/.*\t//'`
      # push
      if [[ ${#tree} > 0 ]]; then
        tree="${tree} $id|$pfx$lpath"
      else
        tree="$id|$pfx$lpath"
      fi
    fi
  done
done