But, then again... The Text::ParseWords module is certainly powerful, but I can't seem to make it handle errors very well. It does not like
"John said "Hello"","again"
So I went back to my earlier method with a better regex

undef $/;
my $infile = (<>);

my @fields = $infile =~ /(".*?"(?=,|\n)|\n)/gms;

my $place = "\n";
foreach (@fields) {
  if ($place ne "\n" && $_ ne "\n") {print  ",";};
  unless ($_ eq "\n") {s/\n//g;};
  print  "$_";
  $place = $_;
}
[download]