The search function now skips '<...>' blocks more reliably and also finds Unicode HTML entities and UTF-8 encoded characters.
