First, ä isn't in that range, so let's look at something that is: —
You won't want to use $utfencode() at all. You can do a $replace of the old character with the unicode version of it (the character isn't gone, it just has a new code) before the check if you need to.
$replace($1-,$chr(151),$chr(8212))
That would be placed on the 7.x client. The 6.35 client should be able to handle just a check of for — or whatever other character.
Or you could just do if ($1- == text $+ $chr(151) $+ text) instead of if($1- == text—text). The downside here is that you would need 2 checks (one for each way). The first example only requires the one check.
I'm sure there are other ways as well.
