With the current code, an encoding failure means we see (and quote)
stuff like \xab\xbc\x07 in the text.  Using the //IGNORE flag allows the
code to ignore broken chars instead of displaying large amounts of junk.

I note that the text containing broken chars is coming from external
sources, so there's nothing that could be done locally to fix it.

-- 
Álvaro Herrera <alvherre@alvh.no-ip.org>