#!/bin/sh

#
# getmails.sh - valid e-mail addresses extractor
#
# Developed by Ondrej Jombik <nepto@platon.sk>
#          and Lubomir Host <rajo@platon.sk>
# Copyright (c) 2001-2003 Platon SDG, http://platon.sk/
# Licensed under terms of GNU General Public License.
# All rights reserved.
#
# $Platon: getmails.sh,v 1.4 2002/06/18 18:24:38 host8 Exp $
#
# Changelog:
#   ??/??/2000 - initial release
#   17/04/2001 - improvements
#   04/02/2002 - import to CVS
#   18/06/2003 - faster version coded in "awk"
#   04/11/2003 - headers update, etc.
#

# Poznamky: (mnohe z nich su out-of-context, su vytiahnute z ineho skriptu)
# adresa sa musi zacinat pismenom (cislom); ak nie, osekame predchadzajuce znaky
# adresa sa musi koncit pismenom; ak nie, vysekame zvysne znaky
# vlavo od znaku '@' je najviac jeden znak '.'
# vpravo od znaku '@' je najmenej jeden znak '.'
# ak vlavo od znaku @ je viac ako 6 cisiel, adresa je neplatna
# domena je max 3 znakova a min 2 znakova a obsahovat moze iba pismena

# filter with awk {{{
use_awk()
{
awk '
BEGIN {
	OFS="\n"
}
// { $0 = tolower($0); }	# convert char. to lowercase

/message-id:/  { next; }	# ignore lines matching this
/content-id:/  { next; }	# ignore lines matching this
/received:/    { next; }	# ignore lines matching this
/in-reply-to:/ { next; }	# ignore lines matching this
/references:/  { next; }	# ignore lines matching this
/^[ \t]*id /   { next; }	# ignore lines matching this

/@/ {	# lines containing @
	fields = split ($0, array, /[^a-z0-9@\.-]+/);
	for (i=1; i<= fields; i++) {	# go over all fields
		if (match(array[i], /^([^\.]+\.[^\.]+|[^\.]+)@(([^\.]+\.)+)[a-z][a-z][a-z]?$/))
			print array[i];
	}
}
' \
| sort -u
}
# }}}

# filter with grep, tr, sed {{{
use_grep()
{
grep -i -v "Message-ID:" \
| grep -i -v "Content-ID:" \
| grep -i -v "Received:" \
| grep -i -v "In-Reply-To:" \
| grep -i -v "References:" \
| tr '[:upper:]' '[:lower:]' \
| tr -c "0-9a-z.@_-" "[\n*]" \
| sed 's/^[^0-9a-z]*//' \
| sed 's/[^a-z]*$//' \
| grep @ \
| grep -E -v '^.*([0-9].*){6,}.*[@].*$' \
| grep -E "^([^\.]+\.[^\.]+|[^\.]+)[@]([^\.]+\.){1,}[a-z][a-z][a-z]?$" \
| sort -u
}
# }}}

# functions for testing speed {{{
test_awk()
{
	for i in 1 2 3 4 5 6 7 8 9 10;
	do
		cat ~/mbox | use_awk > /dev/null;
	done
}

test_grep()
{
	for i in 1 2 3 4 5 6 7 8 9 10;
	do
		cat ~/mbox | use_grep > /dev/null;
	done
}

make_test()
{
	echo "testing use_awk: "; time test_awk;  echo; echo;
	echo "testing use_grep:"; time test_grep; echo; echo;
}

# }}}

# Please specify, which version you want to use
use_awk
#use_grep


###################################################################
# TEST:
# use_awk:   0.56s user 0.16s system 62% cpu 1.149 total
# use_grep:  3.56s user 0.84s system 61% cpu 7.174 total
# RESULT: use_awk is 6.24 fastes as use_grep


# Modeline {{{
# vim:set ts=4:
# vim600:fdm=marker fdl=0 fdc=3 vb t_vb=:
# }}}