lucy24 - 2:09 am on Jan 28, 2013 (gmt 0)

Unhelpful answer #6:

If you had complete up-to-the-minute backups the way people are always telling you to do-- see assorted completely unrelated threads about hacking, server crashes and so on-- this would be a non-problem because all you'd have to do is feed the backup into one of those smart text editors that can spit out all occurrences of the string 'href = "http://([^"]+)'

Or, heck, just run the w3c link checker and save a copy of the results. Sounds as if you're not even asking whether the links are currently valid, just what they are in the first place.