#!/bin/bash
# Copyright 1999-2004 Gentoo Foundation
# Distributed under the terms of the GNU General Public License v2
# Author: Brian Harring <ferringb@gentoo.org>, karltk@gentoo.org originally.
# Rewritten from the old, Perl-based emerge-webrsync script

#-------------------
#initialization
#------------------

f=$(python -c'import portage; print "|".join([portage.settings[x] for x in ("PORTAGE_NICENESS", "GENTOO_MIRRORS", "PORTDIR", "FETCHCOMMAND", "USERLAND", "DISTDIR", "PORTAGE_TMPDIR")])')

IFS='|'
PORTAGE_NICENESS="${f%%|*}";	f="${f#*|}"
GENTOO_MIRRORS="${f%%|*}" ; 		f="${f#*|}"
PORTDIR="${f%%|*}"	 ;		f="${f#*|}"
FETCHCOMMAND="${f%%|*}"	 ;		f="${f#*|}"
USERLAND="${f%%|*}"	 ;		f="${f#*|}"
DISTDIR="${f%%|*}"	 ;		f="${f#*|}"
TMPDIR="${f%%|*}/snapshots"
unset IFS

if [ -z "$NICENESS_PULLED" ]; then
	if [ -n "${PORTAGE_NICENESS}" ]; then
		export NICENESS_PULLED=asdf
		exec nice -n "${PORTAGE_NICENESS}" "$0" "$@"
		echo "failed setting PORTAGE_NICENESS to '$PORTAGE_NICENESS', disabling"
	fi
fi

STATE_DIR="/var/delta-webrsync/"

# hack.  bug 92224
if [ "${FETCHCOMMAND/getdelta.sh}" != "${FETCHCOMMAND}" ]; then
	# evil evil evil evil
	eval "$(grep FETCHCOMMAND /etc/make.globals)"
fi

unset f
unset IFS

MUST_SYNC='1'
unset PUKE_HELP wgetops
for x in $*; do
	if [[ $x == "-u" ]]; then
		MUST_SYNC=''
	elif [[ $x == "-k" ]]; then
		KEEP_OLDIES='asdf'
	elif [[ $x == "-h" ]]; then
		PUKE_HELP=1	
	elif [[ $x == "-v" ]]; then
		wgetops=
	else
		PUKE_HELP=1
		echo "$x isn't a valid arg.  bailing."
	fi
	if [[ -n $PUKE_HELP ]]; then
		echo "-u for upgrade; sync only if new snapshots are found"
		echo "-k for keep; keep old tree snapshots around"
		exit -1
	fi
done

if [[ ! -d $STATE_DIR ]]; then
	echo "$STATE_DIR doesn't exist.  don't have the ability to compensate for compressor differences without it!"
	exit -2
fi

if [[ ! -d $DISTDIR ]] ; then
	mkdir -p $DISTDIR
fi
if [[ ! -d $TMPDIR ]]; then
	mkdir -p $TMPDIR
fi

cd "$DISTDIR"

found=0
if [ "${wgetops-UNSET}" == "unset" ]; then
	#this sucks.  probably better to do 1> /dev/null
	#that said, waiting on the refactoring.
	if [ "${FETCHCOMMAND/wget}" != "${FETCHCOMMAND}" ]; then
		wgetops="-q"
	elif [ "${FETCHCOMMAND/curl}" != "${FETCHCOMMAND}" ]; then
		wgetops="-s -f"
	fi
fi

if type -p md5sum > /dev/null; then
	md5_com='md5sum -c "${MD5_LOC}" &> /dev/null'
elif type -p md5 > /dev/null; then
	md5_com='[ "$(md5 -q ${FILE})" == "$(cut -d \  -f 1 ${MD5_LOC})" ]'
else
	echo "warning, unable to do md5 verification of the snapshot!"
	echo "no suitable md5/md5sum binary was found!"
	md5_com='true'
fi

#---------------
#funcs
#---------------

cleanse_state_dir() {
	[[ ${STATE_DIR:-/} != '/' ]] && rm ${STATE_DIR}/* &> /dev/null;
}

full_version_attempt() {
	local FILE file_exists
	echo "Fetching most recent snapshot"
	declare -i attempts=-1
	while (( $attempts <  40 )) ; do
		unset file_exists
		attempts=$(( attempts + 1 ))

		#this too, sucks.  it works in the interim though.
		if [ "$USERLAND" == "BSD" ] || [ "$USERLAND" == "Darwin" ] ; then
			daysbefore=$(expr $(date +"%s") - 86400 \* $attempts)
			day=$(date -r $daysbefore +"%d")
			month=$(date -r $daysbefore +"%m")
			year=$(date -r $daysbefore +"%Y")
		else
			day=$(date -d "-$attempts day" +"%d")
			month=$(date -d "-$attempts day" +"%m")
			year=$(date -d "-$attempts day" +"%Y")
		fi

		FILE="portage-${year}${month}${day}.tar.bz2"

		echo "Attempting to fetch file dated: ${year}${month}${day}"
		
		got_md5=0

		if [  ! -e "${FILE}.md5sum" ]; then
			fetch_from_mirrors "/snapshots/${FILE}.md5sum" "${FILE}.md5sum"
			got_md5=$?
		else
			file_exists='asdf'
			got_md5=0
		fi

		if [[ $got_md5 != 0 ]]; then
			echo " --- No md5sum present on the mirror. (Not yet available.)"
			continue
		elif [ -s "${FILE}" ]; then
			if verify_md5_file "$FILE"; then
				echo " === snapshot $FILE is correct, using it"
				if [[ -n $MUST_SYNC ]] || [[ -z file_exists ]]; then
					sync_local "${FILE}"
					echo
					echo " === Snapshot has been sync'd"
					echo
				else
					echo 
					echo "skipped sync"
					echo
				fi
				exit 0
			else
				echo "md5 on ${FILE} failed, removing it and starting anew"
				rm "$FILE" &> /dev/null
			fi
		fi
	
		if fetch_from_mirrors "/snapshots/${FILE}" "${FILE}"; then
			if ! verify_md5_file "$FILE"; then
				echo "md5 failed on $FILE"
				rm ${FILE} &> /dev/null
				continue
			else
				sync_local "${FILE}"
				cleanse_state_dir
				echo
				echo " *** Completed websync, please now perform a normal rsync if possible."
				echo "     Update is current as of YYYYMMDD: ${year}${month}${day}"
				echo
				exit 0
			fi
		fi

	done
	exit 1
}


sync_local() {
	local FILE flags
	FILE="$1"
	if [ "${FILE/\/}" == "${FILE}" ]; then
		FILE="${DISTDIR}/${FILE}";
	fi
	
	echo Syncing local tree...
	if type -p tarsync &> /dev/null; then
		echo "apparently you have tarsync installed.  using it."
		if ! tarsync "${FILE}" "${PORTDIR}" -v -s 1 -o portage -g portage -e /distfiles -e /packages -e /local; then
			echo "ok, tarsync failed.  that's teh suck :/"
			exit 6
		fi
	else
		cd ${TMPDIR} || die "couldn't cd to tmpdir, $TMPDIR!?"
		flags="xf"
		if [ "${FILE%.bz2}" != "${FILE}" ]; then
			flags="jxf"
		fi
		if ! tar ${flags} "$FILE"; then
			echo "Tar failed to extract the image. Please review the output."
			echo "Executed command: tar jxf $FILE"
			exit 1
		fi
		# Make sure user and group file ownership is root
		chown -R 0:0 portage
		cd portage
		rsync -av --progress --stats --delete --delete-after \
		--exclude='/distfiles' --exclude='/packages' \
		--exclude='/local' . ${PORTDIR%%/}
		cd ..
		echo "cleaning up"
		rm -rf portage
	fi
	echo "transferring metadata/cache"
	emerge metadata
}

fetch_from_mirrors() {
	local i URI FILE MIRRORS
	if [[ "$#" == 3 ]]; then
		MIRRORS="${3}"
	else
		MIRRORS=$GENTOO_MIRRORS
	fi
	FILE="$2"
	for i in $MIRRORS ; do 
		URI="${i}/${1}"
		if (eval "$FETCHCOMMAND $wgetops") && [ -s "${FILE}" ]; then
			return 0
		else
			rm "${FILE}" &> /dev/null
		fi
	done
	return 1
}

verify_md5_file() {
	local FILE MD5_LOC CUR
	FILE="$1"
	if [[ $# == 2 ]]; then
		MD5_LOC="$2"
	else
		MD5_LOC="$(pwd)/$1.md5sum"
	fi
	if [ "${FILE/*\/}" != "$1" ]; then
		CUR="$(pwd)"
		cd "$(dirname ${FILE})"
		FILE="$(basename ${FILE})"
	fi
	if eval "$md5_com"; then
		[ -n "${CUR}" ] && cd "${CUR}"
		return 0
	else
		[ -n "${CUR}" ] && cd "${CUR}"
		return 1
	fi
}


#--------------------
#inline actual script
#--------------------

if ! type -p patcher &> /dev/null; then
	echo "!!!"
	echo "!!! cannot find patcher, did you emerge dev-util/diffball?"
	echo "!!! lack of patcher == have to do full fetch"
	echo "!!!"
	sleep 10
	full_version_attempt
fi

echo "Looking for available base versions for a delta"

#note we're already in distdir

unset base_version
# portage-snapshots in reverse order.
# icky.
unset dfile
potentials="$(ls -1 portage-200*.tar.bz2 ${STATE_DIR}/portage-200*.tar.bz2 2> /dev/null | sed -e 's:^.*/::' | sort -r)"
for basef in ${potentials}; do
	chksum=''
	found="dar"
	if [ -e "${STATE_DIR}/${basef}.md5sum" ]; then
		chksum="${STATE_DIR}/${basef}.md5sum"
	elif [ -e "${basef}.md5sum" ]; then
		chksum="${DISTDIR}/${basef}.md5sum"
	else
		echo "attempting to get md5sum for $basef"
		if ! fetch_from_mirrors "/snapshots/${basef}.md5sum" "${basef}.md5sum"; then
			echo "can't get md5 for ${basef}"
			continue
		fi
		chksum="${basef}.md5sum"
	fi
	if [ -e "${basef}" ]; then
		dfile="${DISTDIR}/${basef}"
	else
		dfile="${STATE_DIR}/${basef}"
	fi
	if ! verify_md5_file "${dfile}" "${chksum}"; then
		echo "found a stale snapshot.  cleansing"
		rm "${dfile}" &> /dev/null
		rm "${chksum}.md5sum" &> /dev/null
		dar=""
	else
		base_version="${basef}"
		break
	fi
done

#by this point, we either have a base_version, or we don't.
if [[ -z ${base_version} ]]; then
	echo "no base found.  resorting to pulling a full version"
	full_version_attempt
fi

#we have a md5 verified base.  now we get the patch.

base_date="${base_version%.tar.bz2}"
base_date="${base_date#portage-}"
# we now have yyyymmdd

unset patch_dates
if [ "$USERLAND" == "BSD" ] || [ "$USERLAND" == "Darwin" ] ; then
	daysbefore=$(expr $(date +"%s") - 86400 \* $attempts)
	day=$(date -r $daysbefore +"%d")
	month=$(date -r $daysbefore +"%m")
	year=$(date -r $daysbefore +"%Y")
else
	day=$(date -d "-$attempts day" +"%d")
	month=$(date -d "-$attempts day" +"%m")
	year=$(date -d "-$attempts day" +"%Y")
fi

#todays_date="${year}${month}${day}"

next_date() {
	local year day month
	year="${1:0:4}"
	month="${1:4:2}"
	day="${1:6:2}"
#	if [[ "${USERLAND}" == "BSD" ]] || [[ "${USERLAND}" == "Darwin" ]]; then
#	else
		date -d "$year/$month/$day +1 day" +"%Y%m%d"
#	fi
}

patches=''
echo "fetching patches"
fetched='asdf'
while [[ -n ${fetched} ]]; do
	next_day=$(next_date ${base_date})
	# if we can't get a *single* patch or md5, even one missing, do full.
	p="snapshot-${base_date}-${next_day}.patch.bz2"
	if [[ ! -e ${p}.md5sum ]] && ! fetch_from_mirrors "/snapshots/deltas/${p}.md5sum" "${p}.md5sum"; then
		echo "failed fetching ${p}.md5sum"
		fetched=''
		break
	fi
	fetch="yes"
	if [[ -e ${p} ]]; then
		if ! verify_md5_file "${p}"; then
			rm ${p} &> /dev/null
		else
			fetch=""
		fi
	fi
	if [[ -n $fetch ]]; then
		if ! fetch_from_mirrors "/snapshots/deltas/${p}" "${p}"; then
			echo "failed fetching ${p}"
			fetched=''
		fi
	fi
	if [[ -z ${fetched} ]]; then
		break
	fi
	if ! verify_md5_file "${p}"; then
		echo "md5 failed on ${p}"
		fetched=''
		break
	fi
	patches="${patches} ${p}"
	base_date="${next_day}"
done
final_date=${base_date}

if [[ -z $patches ]]; then
	echo "no patches found? up to date?"
	if [[ -n $MUST_SYNC ]]; then
		echo "syncing with existing file"
		sync_local "${dfile}"
	else
		:
	fi
	exit $?
fi

unset got_umd5
#grab the md5 for later usage.
if [[ ! -e portage-${final_date}.tar.bz2.md5sum ]] && ! fetch_from_mirrors "/snapshots/portage-${final_date}.tar.bz2.md5sum" "portage-${final_date}.tar.bz2.md5sum"; then
	echo "warning... couldn't grab the md5sum for ${final_date}.  which is odd"
	echo "thus, bailing (sorry)"
	exit 5
else
	if ! fetch_from_mirrors "/snapshots/portage-${final_date}.tar.bz2.umd5sum" "portage-${final_date}.tar.bz2.umd5sum"; then
		if ! fetch_from_mirrors "/snapshots/portage-${final_date}.tar.bz2.umd5sum" "portage-${final_date}.tar.bz2.umd5sum"; then
			echo "couldn't grab umd5sum (uncompressed md5sum) for ${final_date}."
			echo "can't compensate for bzip2 version differences iow."
		else
			got_umd5=1
		fi
	else
		got_umd5=1
	fi
fi
	
# generate tmp dir.
TEMPDIR=$(mktemp -d /tmp/delta-webrsync-XXXXXX)
# got our patches.
if ! patcher -v "${dfile}" ${patches} "${TEMPDIR}/portage-${final_date}.tar"; then
	echo "reconstruction failed (contact the author with the error from the reconstructor please)"
	rm "${TEMPDIR}/portage-${final_date}.tar"
	rmdir ${TEMPDIR}
	full_version_attempt
fi
verified=0
if [[ -n $got_umd5 ]]; then
	echo "verifying uncompressed md5"
	if ! verify_md5_file "${TEMPDIR}/portage-${final_date}.tar" "${DISTDIR}/portage-${final_date}.tar.bz2.umd5sum"; then
		echo "uncompressed verification failed.  This means either you found a bug in diffball, or something odd is going on"
		echo "with upstream patch generation"
		echo "trying md5sum next, which probably will fail."
	else
		verified="1"
	fi
fi

unset need_last_sync
if [ "$verified" == "1" ]; then
	echo "recompressing. (backgrounding)"
	need_last_sync="dar"
	bzip2 -vk9 "${TEMPDIR}/portage-${final_date}.tar" &

	echo "beginning update to the tree"
	sync_local "${TEMPDIR}/portage-${final_date}.tar"
	echo "doing final md5 stuff"
	wait
	# bzip2 is finished now.
	rm "${TEMPDIR}/portage-${final_date}.tar"
else
	echo "recompressing."
	bzip2 -v9 "${TEMPDIR}/portage-${final_date}.tar.bz2"
fi

echo "verifying generated tarball"

if ! verify_md5_file "${TEMPDIR}/portage-${final_date}.tar.bz2"; then
	if [[ -z $verified ]]; then
		echo "couldn't verify the generated tarball.  bug, most likely."
		exit 5
	fi
	# hokay.  md5 doesn't agree with umd5. bzip2 issue in effect.
	echo "compressed md5 differs, but uncompressed md5 says it right.  bzip2 version incompatability in other words"
	echo "saving the md5"
	if type -p md5sum &> /dev/null; then
		md5sum ${TEMPDIR}/portage-${final_date}.tar.bz2 | sed -e "s:${TEMPDIR}/\?::" > \
			${STATE_DIR}/portage-${final_date}.tar.bz2.md5sum
	elif type -p md5 &> /dev/null; then
		echo "$(md5 -q ${TEMPDIR}/portage-${final_date}.tar.bz2)  portage-${final_date}.tar.bz2" > \
			${STATE_DIR}/portage-${final_date}.tar.bz2.md5sum
	else
		echo "couldn't find either md5 or md5sum.  something is screwed... (bailing, sorry)"
		exit 7
	fi
	mv "${DISTDIR}/portage-${final_date}.tar.bz2.umd5sum" "${TEMPDIR}/portage-${final_date}.tar.bz2" ${STATE_DIR}/
	rmdir ${TEMPDIR}
	dfile="${STATE_DIR}/portage-${final_date}.tar.bz2"
else
	dfile="${DISTDIR}/portage-${final_date}.tar.bz2"
	mv "${TEMPDIR}/portage-${final_date}.tar.bz2" ${DISTDIR}/
fi

if [ -z "${need_last_sync}" ]; then
	echo "beginning update to the tree"
	sync_local "${dfile}"
fi

if [[ -z $KEEP_OLDIES ]]; then
	echo "cleansing"
	for x in $potentials; do
		echo "removing ${x}"
		rm "${DISTDIR}/${x}" "${DISTDIR}/${x}.md5sum" &> /dev/null
		rm "${STATE_DIR}/${x}" "${STATE_DIR}/${x}.md5sum" "${STATE_DIR}/${x}.umd5sum" &> /dev/null
	done
fi
echo "done."