This commit switches back to using the regexp engine for
case-insensitive literal string matches.
This is slower, but at least case-insensitive matches for string
literals will function now. The code is a tiny bit shorter and simpler
too.
Given the aim of the tool is to be useful for ad-hoc searches,
efficiency isn't the concern but rather just getting the job done with
the minimum of fuss / unexpected behaviour.
Laurence Withers