Markus Bertheau <twanger(at)bluetwanger(dot)de> writes:
> pg's regex engine does have one
> shortcoming though: it doesn't know UTF-8.
Sure it does. We borrowed it from Tcl, remember?
The "character class" stuff is not locale-aware at the moment,
which is something that ought to get fixed eventually, but claiming
it doesn't handle UTF8 at all is simply wrong.
regards, tom lane