I want to get rid of anything between < and > from form data unless it is a
hyperlink.
Taking from this form parser (asterisks showing the line I want to modify):

sub Parse_Form {
 if ($ENV{'REQUEST_METHOD'} eq 'GET') {

    $ENV{'QUERY_STRING'});
 } elsif ($ENV{'REQUEST_METHOD'} eq 'POST') {
  read (STDIN, $buffer,
    $ENV{'CONTENT_LENGTH'});

  if ($ENV{'QUERY_STRING'}) {


   }
 } else {
  print "Content-type: text/html\n\n";
  print "<P>Use Post or Get";
 }

  ($key, $value) = split (/=/, $pair);
  $key =~ tr/+/ /;
  $key =~ s/%([a-fA-F0-9][a-fA-F0-9])/
    pack("C", hex($1))/eg;
  $value =~ tr/+/ /;
  $value =~ s/%([a-fA-F0-9][a-fA-F0-9])/
    pack("C", hex($1))/eg;

  $value =~s/<!--(.|\n)*-->//g;   #
*****************************************************

  if ($formdata{$key}) {
   $formdata{$key} .= ", $value";
  } else {
   $formdata{$key} = $value;
  }
 }

I want to change the line:  $value =~ s/<!--(.|\n)*-->//g; (shown by
asterisks) to something like the following:

$value =~ s/<[^a href]*?>//gi;

The only problem is that the substitution is looking for any one of the
letters: a, space, h, r, e, and f.  It won't allow me to substitute a
negated string of this form.  Is there a way to negate 'a href' without the
brackets?  How else might I go about this?  The goal is to remove any HTML
and SSI except for hyperlinks that a person may enter into a form.  I know I
can see if I can match something I don't want in the form data and tell the
user not to do that but I would rather have the script remove it.

Thanks for any help you can offer.

-Mike Morgan

require 5.003;
use strict;

my $yes = qq(blah <img src="http://kistler.discordia.ch/x.jpg"> blah);
my $no  = qq(blah <a href="http://kistler.discordia.ch/">Unix</a> blah);

for my $try ( $yes, $no ){
        $try =~ s(<(?!a href|/a).*?>)
                 ( "removed here" )gi;
        print $try,"\n";