#! @PERL@

use warnings;

# RFCs important for this script:
#
# RFC 2822 - Internet Message Format
# RFC 2047 - MIME (Multipurpose Internet Mail Extensions) Part Three:
#            Message Header Extensions for Non-ASCII Text

use Getopt::Long;
use strict;

# This ugly trick lets the script work even if gettext support is missing.
# We did so because Locale::gettext doesn't ship with the standard perl
# distribution.
BEGIN {
    if (eval { require Locale::gettext }) {
	import Locale::gettext;
	require POSIX;
	import POSIX, qw(setlocale);
    } else {
	eval '
	    use constant LC_MESSAGES => 0;
	    sub setlocale($$) { }
	    sub bindtextdomain($$) { }
	    sub textdomain($) { }
	    sub gettext($) { shift }
	'
    }
}

setlocale(LC_MESSAGES, "");
bindtextdomain("quilt", $ENV{'STAGING_DIR_HOST'} ? $ENV{'STAGING_DIR_HOST'} . '/share/locale' : "@LOCALEDIR@");
textdomain("quilt");

sub _($) {
    return gettext(shift);
}

my $special = '()<>\[\]:;@\\,"';  # special characters
my $special_dot = "$special.";  # special characters + dot

sub check_recipient($);

my (%append_name, %append_value, $remove_empty_headers, %remove_header,
    %extract_recipients_from, %replace_name, %replace_value, $charset);
GetOptions('add-recipient=s%' =>
    sub {
	$append_name{lc $_[1]} = $_[1];
	$append_value{lc $_[1]} .= ",\n " . $_[2];
    },
    'add-good-recipient=s%' =>
    sub {
	eval { check_recipient($_[2]) };
	if ($@) {
		chomp $@;
		print STDERR "$@; skipping\n";
	} else {
		$append_name{lc $_[1]} = $_[1];
		$append_value{lc $_[1]} .= ",\n " . $_[2];
	}
    },
    'remove-header=s' => sub { $remove_header{lc $_[1]}++ },
    'remove-empty-headers' => \$remove_empty_headers,
    'replace-header=s%' =>
    sub {
	$replace_name{lc $_[1]} = $_[1];
	$replace_value{lc $_[1]} = $_[2];
    },
    'extract-recipients=s' => sub { $extract_recipients_from{lc $_[1]} = 1 },
    'charset=s' => \$charset)
    or exit 1;
my %recipient_headers = map {lc $_ => 1} (@ARGV, keys %append_name);

sub encode_header($) {
    my ($word) = @_;
    $word =~ s{[^\t\41-\76\100-\176]}{sprintf "=%02X", ord($&)}ge;
    return "=?$charset?q?$word?=";
}

# Check for a valid display name
sub check_display_name($) {
    my ($display) = @_;

    if ($display =~ /^"((?:[^"\\]|\\[^\n\r])*)"/) {
	my $quoted = $1;
	if ($quoted =~ /[^\t\40-\176]/) {
	    $display = $quoted;
	    $display =~ s/\\//;
	    return encode_header($display);
	}
    } else {
	local $_ = $display;
	# The value is not (properly) quoted. Check for invalid characters.
	while (/\(/ or /\)/) {
	    die sprintf(
_("Display name `%s' contains unpaired parentheses\n"), $display)
		unless s/\(([^()]*)\)/$1/;
	}
	if ($display =~ /[^\t\40-\176]/ || $display =~ /[$special_dot]/) {
	    if ($display =~ /[^\1-\10\13\14\16-\37\40\41\43-\133\135-\177]/) {
		return encode_header($display);
	    } elsif ($display =~ /[$special_dot]/) {
		return "\"$display\"";
	    }
	}
    }
    return $display;
}

# Check for a valid delivery address
sub check_delivery_address($) {
    my ($deliver) = @_;

    die sprintf(_("Delivery address `%s' is invalid\n"), $deliver)
	if $deliver =~ /[ \t]/ or
	   $deliver =~ /[^ \t\40-\176]/ or
	   $deliver !~ /^[^$special]+@(\[?)[^$special_dot]+(?:\.[^$special_dot]+)*(\]?)$/ or
	   (!$1) != (!$2);
    return $deliver;
}

sub check_recipient($) {
    my ($recipient) = @_;

    if ($recipient =~ /^(.*?)\s*<(.+)>$/) {
	my $deliver = check_delivery_address($2);
	return ( check_display_name($1) . " <" . $deliver . ">", $deliver );
    } elsif ($recipient =~ /^(\S*)\s*\((.*)\)$/) {
	my $deliver = check_delivery_address($1);
	return ( $deliver . " (" . check_display_name($2) . ")", $deliver );
    } else {
	my $deliver = check_delivery_address($recipient);
	return ( $deliver, $deliver );
    }
}

sub split_recipients($) {
    my ($recipients) = @_;
    my @list = ();

    while ($recipients !~ /^\s*$/) {
	my $recipient;
	if ($recipients =~ s/^\s*,?\s*((?:"(?:[^"]+)"|[^",])*)//) {
	    $recipient = $1;
	} else {
	    $recipient = $recipients;
	    $recipients = "";
	}
	$recipient =~ s/\s*$//;
	push @list, $recipient;
    }
    return @list;
}

my %recipients;
sub process_header($) {
    local ($_) = @_;
    my ($name, $value);

    return unless defined $_;
    unless (($name, $value) = /^([\41-\176]+):\s*(.*)/s) {
	print;
	return
    }
    if (%extract_recipients_from) {
	if (exists $extract_recipients_from{lc $name}) {
	    $value =~ s/^\s*//;  $value =~ s/\s*$//;
	    foreach my $recipient (split_recipients($value)) {
		    my $deliver;
		    ($recipient, $deliver) = check_recipient($recipient);
		    print "$deliver\n";
	    }
	}
	return;
    }
    return if exists $remove_header{lc $name};
    if (exists $replace_name{lc $name}) {
	    if (exists $replace_value{lc $name}) {
		print "$replace_name{lc $name}: $replace_value{lc $name}\n";
		delete $replace_value{lc $name};
	    }
	    return;
    }
    if (exists $recipient_headers{lc $1}) {
	if (exists $append_name{lc $name}) {
	    $value .= $append_value{lc $name};
	    delete $append_name{lc $name};
	}
	my @recipients;
	# This is a recipients field. Split out all the recipients and
	# check the addresses. Suppress duplicate recipients.
	foreach my $recipient (split_recipients($value)) {
	    my $deliver;
	    ($recipient, $deliver) = check_recipient($recipient);
	    unless (exists $recipients{$deliver}) {
		push @recipients, $recipient;
		$recipients{$deliver} = $deliver;
	    }
	}
	print "$name: ", join(",\n ", @recipients), "\n"
	    if @recipients || !$remove_empty_headers;
    } else {
	    print if $value ne "" || !$remove_empty_headers;
    }
}

my $header;
while (<STDIN>) {
    last if (/^$/);
    if (/^\S/) {
	process_header $header;
	undef $header;
    }
    $header .= $_;
}
process_header $header;
foreach my $name (keys %append_name) {
    process_header $append_name{$name} . ': ' . $append_value{$name};
}
unless (%extract_recipients_from) {
    # Copy the message body to standard output
    # FIXME check for 7-bit clean, else assume $charset
    # FIXME if UTF-8, check for invalid characters!
    # FIXME must make sure that all messages are written in
    #       either 7-bit or $charset => mbox !!!

    # Content-Transfer-Encoding: 7bit
    # Content-Transfer-Encoding: 8bit
    # Content-Type: text/plain; charset=ISO-8859-15
    # Content-Type: text/plain; charset=UTF-8
    undef $/;
    print "\n", <STDIN>;
}