#!/bin/sh

#
# distclean.sh: Clean out distfiles that are not referenced by
#               any installed packages.
#

# $Header: /var/cvsroot/scripts/distclean.sh,v 1.4 2005/05/22 14:12:25 mark Exp $
#
# Copyright (C) 2005 Mark Rosenstand <mark@borkware.net>
#
#  This program is free software; you can redistribute it and/or modify
#  it under the terms of the GNU General Public License as published by
#  the Free Software Foundation; either version 2 of the License, or
#  (at your option) any later version.
#
#  This program is distributed in the hope that it will be useful,
#  but WITHOUT ANY WARRANTY; without even the implied warranty of
#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#  GNU General Public License for more details.
#
# See http://www.gnu.org/licenses/gpl.txt for details.
#

# Try to inherit DISTDIR from portage configs.
[ -e /etc/make.globals ] && . /etc/make.globals
[ -e /etc/make.conf ] && . /etc/make.conf

# Fall back to this in case inheriting from portage fails.
[ -z ${DISTDIR} ] && DISTDIR="/usr/portage/distfiles"

# Constant value VDB_PATH from portage.
PKG_DB_DIR="/var/db/pkg"

# Check if /dev/shm exists and is writable, fall back to /tmp if not.
[ -d /dev/shm -a -w /dev/shm ] \
	&& TMP_FILELIST="/dev/shm/distclean.$$" \
	|| TMP_FILELIST="/tmp/distclean.$$"

usage() {
	echo "
distclean.sh: Clean out distfiles that are not referenced by
              any installed packages.

  The advantage of doing this is that you can keep a copy of the source
  code for installed packages without wasting too much disc space.

  - \"But, why keep the source code around?\"

  Every time you re-compile a package, you need it. This could be
  because you changed a USE flag or the maintainer bumped its revision.
  This happens quite often; for instance, sys-kernel/gentoo-sources
  typically bumps its revision 10 times for each release.

  Deleting your distfiles will mean that the next time you re-compile
  a package, you will have to re-fetch the entire source code, even
  if the change was to apply a 2 kB patch or just change something in
  the build instructions. That's 40 MB for sys-kernel/gentoo-sources.
  This is an awful waste of bandwidth, time and mirror resources.

  - \"OK, so I won't delete anything in /usr/portage/distfiles.\"

  /usr/portage/distfiles will steadily grow as you update packages,
  try out new applications, etc. I wrote this utility because I had
  5 GB of distfiles after 6 months of weekly emerge -uD world'ing.
  Those were reduced to 900 MB the first time I ran this utility.

  I hope that this script will help others spare some disc space
  *and* not waste mirror resources.

                                -- Mark Rosenstand <mark@borkware.net>


Usage:

    $0 [options]

Options:

  -h, --help
      This message.

  -d, --debug
      Enable output useful for debugging.

  -v, --verbose
      Enable more verbose output.

  -q, --quiet
      Don't print any non-error messages.

  -i, --interactive
      Ask before trying to delete files.

  -p, --pretend
      Don't actually delete the files.

  WARNING: Currently, the last option will always take precedence,
           so '$0 -q -v' will *not* be quiet.
"
}

die() {
	echo
	echo "  Error: $1" >/dev/stderr
	echo
	exit 1
}

# Read parameters.
while [ $# -gt 0 ]; do
	case "$1" in
		-d|--debug)
			DEBUG=1
		;;
		-v|--verbose)
			VERBOSE=1
		;;
		-q|--quiet)
			QUIET=1
		;;
		-p|--pretend)
			PRETEND=1
		;;
		-i|--interactive)
			INTERACTIVE=1
		;;
		-h|--help)
			usage
			exit 0
		;;
		*)
			usage
			die "$1 is not a valid option."
		;;
	esac
	shift
done

check_param_set() {
	[ ${1} ] && echo "true" || echo "false"
}

if [ ${DEBUG} ]; then
	echo
	echo " Configuration"
	echo "========================================================"
	echo "  distfiles directory:         ${DISTDIR}"
	echo "  package database directory:  ${PKG_DB_DIR}"
	echo "  temporary list of files:     ${TMP_FILELIST}"
	echo
	echo " Parameters"
	echo "========================================================"
	echo -n "  --debug:                     "; check_param_set ${DEBUG}
	echo -n "  --verbose:                   "; check_param_set ${VERBOSE}
	echo -n "  --quiet:                     "; check_param_set ${QUIET}
	echo -n "  --interactive:               "; check_param_set ${INTERACTIVE}
	echo -n "  --pretend:                   "; check_param_set ${PRETEND}
	echo
fi

# Clean out old file list.
if [ -f ${TMP_FILELIST} ]; then
	rm -f ${TMP_FILELIST} || die "Can't remove ${TMP_FILELIST}"
fi

[ -z ${QUIET} ]	\
	&& echo "Generating list of distfiles used by installed packages..."

for CATEGORY in `ls ${PKG_DB_DIR}`; do
	[ ${VERBOSE} ] && echo "${CATEGORY}/"
	for PKG in `ls ${PKG_DB_DIR}/${CATEGORY}`; do
		[ ${VERBOSE} ] && echo "    ${PKG}:"
		FILES=`bzcat ${PKG_DB_DIR}/${CATEGORY}/${PKG}/environment.bz2 \
			|grep "^A=" |sed -e 's/A=//;' -e "s/'//g"`
		for FILE in ${FILES}; do
			[ ${VERBOSE} ] && echo "        ${FILE}"
			echo "${FILE}" >>${TMP_FILELIST} \
				|| die "Can't write to ${TMP_FILELIST}"
		done
	done
done

[ -z ${QUIET} ] \
	&& echo "Checking files in ${DISTDIR} against the list..."

for DISTFILE in `ls ${DISTDIR}`; do
	# Check if the distfile is in the list.
	if grep -q "^${DISTFILE}$" ${TMP_FILELIST}; then
		# Print a message if we're being --verbose.
		[ ${VERBOSE} ] && echo "Keeping ${DISTFILE}"
	else
		# Print a message if we're not --quiet.
		[ -z ${QUIET} ] && echo "Removing ${DISTFILE}"
		while true; do
			# Skip all steps if we --pretend.
			[ ${PRETEND} ] && break
			# Prompt user if we're --interactive.
			if [ ${INTERACTIVE} ]; then
				echo -n "Delete ${DISTFILE}? [Y/n] "
				read REPLY
				case "${REPLY}" in
					yes|y|Y)
						rm -rf "${DISTDIR}/${DISTFILE}"
						break
					;;
					no|n|N)
						break
					;;
					*)
						echo "Please choose 'yes' or 'no'."
						continue
					;;
				esac
			else
				# Just delete the file if neither of the above.
				rm -rf "${DISTDIR}/${DISTFILE}"
				break
			fi
		done
	fi
done

# Be a good kid and clean up after playing.
# Unless we're --debug'ging, of course.
[ -z ${DEBUG} ] \
	&& rm -f ${TMP_FILELIST} \
	|| echo "List of referenced distfiles saved in ${TMP_FILELIST}." \

exit 0