Find and group differences between two XML trees.

changex.sh

#!/bin/bash

if [[ -f "/usr/bin/shasum" ]]; then
  SHASUM='/usr/bin/shasum'
else SHASUM='/usr/bin/sha1sum'
fi

if [[ -n "$XML1" ]] && [[ -n "$XML2" ]] && [[ "$XML1" != "$XML2" ]]; then

  profiles=$(diff $XML1 $XML2 --brief | grep -v '\.dep' | grep -v '\.json\.gz' | grep -v profiles-info.xml | grep -v "Only in" | awk '{ print $2 }' | sed "s/^.*\///g")

  count=0

  for p in $profiles; do
    count=$(($count + 1))
  done

  pos=0

  echo -n "Examining profiles, please wait     " 1>&2
  exemplars="$(for p in $profiles; do
    pos=$(($pos + 1))
    echo -n "$p: "
    perc="$((($pos * 100) / $count))"
    perc=$(printf "%3i%%" $perc)
    echo -en "\b\b\b\b$perc" 1>&2
    diff -w $XML1/$p $XML2/$p | grep -v "^[0-9]" | sort | $SHASUM
  done | rev | sort | rev | uniq -f 1 -c | cut -d : -f 1 | awk '{ print $2","$1 }')"

  echo "Completed."

  for x in $exemplars; do
    p=$(echo $x | cut -d , -f 1)
    n=$(echo $x | cut -d , -f 2)
    echo "---- $p (exemplar for $n profiles) ----"
    colordiff -w $XML2/$p $XML1/$p $@
  done

else

  echo "Please set environment variables XML1 and XML2 to point to the locations of two different XML trees."
  echo "XML1=$XML1"
  echo "XML2=$XML2"

fi