The Wikipedia clone site http://www.splammer.com/ appears to me to be copying Wikipedia content in real time, apparently by reading the raw Wikitext, rather than running from a dump.

Could someone take a look to see whether this is the case?

-- Neil