#!/bin/bash

WARN_ON_SIZE=500000
ERROR_ON_SIZE=1000000
EXIT=0
FIRST_COMMIT_HASH=$1
LAST_COMMIT_HASH=$2

# https://en.wikipedia.org/wiki/ANSI_escape_code#3-bit_and_4-bit
function echo_red () { printf "\e[0;31m$1\e[m\n"; }
function echo_yellow () { printf "\e[0;34m$1\e[m\n"; }
function echo_grey () { printf "\e[0;90m$1\e[m\n"; }

while read file
do
    echo "Checking ${file}..."
    # git ls-tree will output:
    # (file mode) (file type) (blob hash)<TAB>(file name)
    # we're interested in the hash to pull the file's size using cat-file
    if [[ -z "${LAST_COMMIT_HASH}" ]]; then
        # manually check in local ci runs to account for uncommitted files
        filesize=`wc -c < ${file}`
    else
        hash=`git ls-tree ${LAST_COMMIT_HASH} "${file}" | awk -F ' ' '{ print $3 }'`
        filesize=`git cat-file -s ${hash} 2>/dev/null`
    fi
    if [[ ${filesize} -ge ${ERROR_ON_SIZE} ]]; then
        printf "$( echo_red 'Error:' ) The size of \"${file}\" exceeds 1MB. Compress it to optimise performance.\n"
        EXIT=1
    elif [[ ${filesize} -ge ${WARN_ON_SIZE} ]]; then
        printf "$( echo_yellow 'Warning:' ) The size of \"${file}\" exceeds 500kB. Consider compressing it to optimise performance.\n"
    fi
done < <(git diff --numstat --no-renames --diff-filter=d ${FIRST_COMMIT_HASH}^ ${LAST_COMMIT_HASH} | awk '/^-\t-\t/ { print $3 }')
# git diff --numstat will output -<TAB>-<TAB>$filename for blobs

if [[ ${EXIT} -eq 0 ]]; then
    printf "$( echo_grey 'Notice:' ) No excessively large files detected\n"
fi

exit ${EXIT}