That is not even remotely like valid XML, but anyway...

You can try this (ignoring CDATA sections, but including other weird stuff, like the XML declaration):

@tags = grep defined, $xml =~ /<!--.*?-->|(<(?>[^"'>]+|'[^']*'|"[^"]*"
+)*>)/sg;
[download]

And of course, you'll have some cleaning up to do now, because the output is very coarse.

As a first step into parsing this content, you can use the above regexp in split:

@tokens = split /<!--.*?-->|(<(?>[^"'>]+|'[^']*'|"[^"]*")*>)/s, $xml;
[download]

which will return a list of text and tags. Comments will be thrown away.