# Extract \Title entries from PDFs embedded in the challenge file
# Fairly simple and ad hoc (based on PDF files), but the mold should be 
# obvious: get some unique identifier and search for it. Here we get it; the
# other program searches for it, and then the directory intersect program throws
# away the false positives.

#my $fn = "e76.pdf";

if ($#ARGV < 0) {
	die "Usage: extract_titles.pl [file to extract PDF titles from]\n";
}

my $fn = $ARGV[0];

print "\n";

open (CHALLFILE, $fn) || die "Can't open file!";

while ($potential = <CHALLFILE>) {
	@divided_titles = grep(/\/Title/, $potential);
	@potential_titles = split(/\//, join("\n", @divided_titles));
	@twice_titles = grep(/Title/, @potential_titles);
	#@divided_titles = grep(/\/Title/, $potential_titles);
#	@titles_this = grep(/Title/, split(/\//, grep(/\/Title/, $potential)));
	push (@titles, @twice_titles);
	# Split from \Title to endline
}

# Clean up
foreach $fh (@titles) {
	# Remove binary stuff before the title
	$fh =~ s/.*Title/Title/g;
	# Remove other elements after
	$fh =~ s/\).*/\)/g;
	# Be sure control chars are gone
	$fh =~ s/\\[[:cntrl:]]//g;
	$fh =~ s/[[:cntrl:]]//g;

	#Unescape octal
	$fh =~ s/\\(0[0-7]+)/chr(oct($1))/esg;
	#Prune unicode
	$fh =~ s/\000//g;

	#And finally remove the title elements.
	$fh =~s/^Title\(//g;
	$fh =~s/^Title \(//g;
	$fh =~s/\)$//g;
}

print join("\n", @titles), "\n";

close(CHALLFILE);