Preg_match_all not stopping where it should be
run the regular expression in U-ngready mode
preg_match_all('#<cite>(.+)</cite>#siU
As in Darhazer's answer you can turn on ungreedy mode for the whole regex using the U pattern modifier, or just make the pattern itself ungreedy (or lazy) by following it with a ?:
preg_match_all('#<cite>(.+?)</cite>#si', ...