Isn't that impossible? You can redirect to a page (under your control) that redirects to itself, if your goal is to get the user stuck in a redirect loop. Until their browser decides enough is enough, that is.

You can have the service test the link for redirects (using HEAD requests), and have a max depth to be safe. Then if a URL occurs twice, or the max depth is reached, mark as invalid.

"browser decides enough is enough"

IIRC from reading the sources, both FF and Chrome had a limit of 30 hardcoded - after this many redirects they would give up and say "uh, it's a redirect loop".

Nice try, but Chrome is smarter than that. "This webpage has a redirect loop. The webpage at http://gadaf.fi/5j has resulted in too many redirects..."