what is wrong with my regex pattern?

Anon

My Code:
<?php

$fullDoc = "<HEAD><TITLE>TEST DOC</TITLE></HEAD><BODY><UL><LI><A HREF='mike.html'><H3>Test Link</H3></A></LI></UL></BODY>";
$pat = "/<A(.+)HREF=(?:\"')(?:\"')>/Ui";
$repl = "<A HREF=\"http://www.myserver.com/parse.php?url=$url&link=\2\" \1 \3>";
$new=preg_replace($pat, $repl, $fullDoc);
print $new;

?>
Ok, I want to find all Anchor links are replace them with $repl. Simpple enough. I need to account for all differnt anchor conventions though, which would be: Tags between <A and HREF, HREF='link', HREF="link", HREF=link, and tags after the HREF (i.e. TARGET="top").
It all works like a CHARM, except it does NOT match if the are not quotes in the href, example. <A TARGET="TOP" HREF"=link.html" NAME=Link1> works, but <A TARGET="_TOP" HREF=link.html NAME=Link1> with out the quotes doesn't work... Anyone help me here?
Thanks ALOT!!!
mike

Anon

Well, the forum window intepreted half of my HTML coding, but you can get the point, the anchor tag is all that is important and the pattern, so... sorry

Anon

change:

/<A(.+)HREF=(?:\"')(?:\"')>/Ui

/<A(.+)HREF=['"]?(.+)['"]?(.*)>/Ui

I'm not sure why you had non-capturing parens inside character-classes - They won't work.

Anon

That expression does not work, it's hard to say why, but... the source code looks like this: <A HREF="http://www.myserver.com/parse.php?url=&link=t" est.html>Test Link</A> I can't explain why, if there is a tick or quote, then it would like like yadda?url=&link="" test.html>
Mike