use HTML::Entities qw( encode_entities );

my $url = 'http://www.somedomain.com/index.cgi?page=home&var=1&no=2&so
+=forth&so=on';

my $h_url = encode_entities($url);
print(qq{...<a href="$h_url">...</a>...};
[download]

Using Template-Toolkit, you'd use something like

...<a href="[% url | html %]">...</a>...
[download]

In short:

use HTML::Entities;
encode_entities($_all_page_content);
[download]

gmargo pointed to the right direction in his reply:

$ perl -we 'use CGI; $url = "http://www.somedomain.com/index.cgi?page=
+home&var=1&no=2&so=forth&so=on"; print qq[<a href="], CGI::escapeHTML
+($url), qq[">description</a>\n];'
<a href="http://www.somedomain.com/index.cgi?page=home&amp;var=1&amp;n
+o=2&amp;so=forth&amp;so=on">description</a>
[download]

It's of course easier to escape the ampersands before you compose the whole text of the page in a variable, because in the latter case it's hard to find out what exactly to escape, and if there are unescaped quotes or angle brackets or ampersands that look like they start valid xml escapes, you might not even be able to parse the results out.

This is totally overkill for what you want but it's a nice approach for converting HTML to XHTML and that's the subtext of your original question.

use strict;
use warnings;
use XML::LibXML;

my $raw = do { local $/; <DATA> };

my $parser = XML::LibXML->new;
$parser->recover_silently(1);

my $doc = $parser
    ->parse_html_string("<div>$raw</div>");

my $wrapper = [ $doc->findnodes("//body/div") ]->[0];

print $_->serialize(1)
    for $wrapper->childNodes;

exit 0;

__DATA__
All content is in a variable like this...<br>
<br>
<a href="http://www.somedomain.com/index.cgi?page=home&var=1&no=2&so=f
+orth&so=on">
[download]

You end up with-

What does this have to do with XML? He specifically mentioned HTML where ampersand needs to be escaped as well. Especially when you consider that the semi-colon is optional in HTML.

http://www.somedomain.com/index.cgi?page=home&quot=1
[download]

is the same as

http://www.somedomain.com/index.cgi?page=home"=1
[download]

The OP asked how to get &s to &s in an existing snippet of HTML. My solution does that and it does it without saying start over and do it from scratch which often isn't an option.

It also sounded like it might be an XY question and the real issue was a need to do HTML->XHTML since the vanilla HTML world isn't usually concerned whether docs validate or not.

See also the escapeHTML() routine provided by the CGI module.