Diff directories to see which files are contributing to file size regressions

diff-city.sh

#!/bin/bash

set -e

if [[ -z "$1" || -z "$2" ]]; then
	echo "usage: ./diff-city.sh new old"
	exit 1
fi

old=$1
new=$2
mkdir -p /tmp/diff-city/
cd $old
find . -type f -exec du -ka {} + > /tmp/diff-city/du-old
cd ..
cd $new
find . -type f -exec du -ka {} + > /tmp/diff-city/du-new
cd ..

rm -f /tmp/diff-city/report.tmp

old_total_lines=$(wc -l < /tmp/diff-city/du-old | xargs)
count=0
while read line; do
	((count+=1))
	echo "processing $old manifest: $count / $old_total_lines"
	IFS=$'\t' read old_size file <<< $line
	IFS=$'\t' read new_size trash <<< $(cat /tmp/diff-city/du-new | grep "\t$file\$")

	if [ -z "$new_size" ]; then
		echo "[$file] -${old_size}k (DELETED)" >> /tmp/diff-city/report.tmp
		continue
	fi

	size_diff=$(($new_size-$old_size))

	if [ "$new_size" -eq "$old_size" ]; then
		continue
	fi

	echo "[$file] ${size_diff}k" >> /tmp/diff-city/report.tmp
done < /tmp/diff-city/du-old

new_total_lines=$(wc -l < /tmp/diff-city/du-new | xargs)
count=0
while read line; do
	((count+=1))
	echo "processing $new manifest: $count / $new_total_lines"
	IFS=$'\t' read size file <<< $line
	IFS=$'\t' read exists_before <<< $(cat /tmp/diff-city/du-old | { grep "\t$file\$" || true; })

	if [ -z "$exists_before" ]; then
		echo "[$file] ${size}k (NEW)" >> /tmp/diff-city/report.tmp
	fi
done < /tmp/diff-city/du-new

sort -k2 -n -r /tmp/diff-city/report.tmp