Update Opera's urlfilter.ini from secure.fanboy.co.nz while preserving custom entries

update_urlfilter.ini.pl

#!/usr/bin/perl -w
use LWP::Simple;
use Storable;
use strict;

# Configuration
# You can define multiple urlfilter.ini lists to be downloaded
my @remotes = qw(
	http://secure.fanboy.co.nz/adblock/opera/urlfilter.ini
);

# File in which to store a copy of the remote list
# This is required for the script to distinguish between removed
# entries and local entries
my $remote_store_filename = 'urlfilter.ini.plinfo';

##############################################################

if(`pgrep -clf opera` > 0) {
	die "Opera is running. Not updating.";
}

# Download new urlfilter lists
my $remote_urlfilter_data;
for(@remotes) {
	my $a_remote_urlfilter_data = get($_);
	if($a_remote_urlfilter_data eq undef) {
		warn "Failed to download urlfilter.ini from " . $_;
	}
	$a_remote_urlfilter_data =~ s/^.+\[exclude\]\s+//s;
	$a_remote_urlfilter_data =~ s/^#.+\r?\n//m;
	chomp $a_remote_urlfilter_data;
	$remote_urlfilter_data .= $a_remote_urlfilter_data . "\n";
}
my %remote_urlfilter = map {
	s/\s+$//;
	if(m/^"([^"]+)"(=UUID:.+)?$/) {
		lc $1 => $2 || 1;
	}
	elsif(m/^([^=]+)(=UUID:.+)?$/) {
		lc $1 => $2 || 1;
	}
} split(/\r?\n|\r(?!\n)/, $remote_urlfilter_data);

my %remote_urlfilter_old;

if($remote_store_filename) {
	# Load stored copy of downloaded list
	%remote_urlfilter_old = -e $remote_store_filename ? %{retrieve($remote_store_filename)} : {};
	# And store a copy of the new list
	store \%remote_urlfilter, $remote_store_filename;
}

# Load old urlfilter list
my $file_start = "";
open URLFILTER, "<urlfilter.ini";
while(1) {
	my $line = <URLFILTER>;
	$file_start .= $line;
	last if($line =~ m/^\[exclude\]/);
	last if eof;
}
unless($file_start) {
	$file_start = <<EOF
Opera Preferences version 2.1
; Do not edit this file while Opera is running
; This file is stored in UTF-8 encoding

[prefs]
prioritize excludelist=1

[include]
*=UUID:1924F7C05D4311E0891A8B052607A998

EOF
}
if(eof) {
	$file_start .= "[exclude]\n";
}
my %urlfilter = map {
	s/\s+$//;
	if(m/^"([^"]+)"(=UUID:.+)?$/) {
		lc $1 => $2 || 1;
	}
	elsif(m/^([^=]+)(=UUID:.+)?$/) {
		lc $1 => $2 || 1;
	}
} <URLFILTER>;
close URLFILTER;

# Create an updated list
my ($local, $added, $removed);
$local = $added = $removed = 0;

# Remove remotely removed entries from local list
# This is only possible if the list was sync'ed before, since we can't tell
# local stuff apart from remotely removed..
for (keys %urlfilter) {
	unless(exists($remote_urlfilter_old{$_})) {
		# Must be present in old remote list, elsewise it's a local rule
		$local++;
		next;
	}
	next if(exists($remote_urlfilter{$_})); # But not present in new remote list
	$removed++;
	undef $urlfilter{$_};
}

# Insert new entries
for (keys %remote_urlfilter) {
	next if(m/^\s*$/);
	next if(exists($urlfilter{$_}));
	$added++;
	$urlfilter{$_} = 1;
}

# Write an updated list
if($added + $removed > 0) {
	print "Update complete. " . $added . " entries added, " . $removed . " removed, " . $local . " local rules kept.\n";
	open URLFILTER, ">urlfilter.ini";
	print URLFILTER $file_start;
	for (keys %urlfilter) {
		next if m/^\s*$/;
		next unless $urlfilter{$_};
		if(m/=/) {
			print URLFILTER '"' . $_ . '"';
		}
		else {
			print URLFILTER $_;
		}
		print URLFILTER $urlfilter{$_} if($urlfilter{$_} ne 1);
		print URLFILTER "\n";
	}
	close URLFILTER;
}
else {
	print "urlfilter.ini is up to date\n";
}