Andrew Dunstan <andrew@dunslane.net> writes:
> Off the top of my head I would have thought there was a good case for
> raising an error on chr(0). Aren't null bytes forbidden in text values?
They're not supported, but we don't make any strenuous efforts to
prevent them. A watertight prohibition would require extra effort in a
lot of places, not only chr(). The string literal parser and text_recv
and friends come to mind immediately; there are probably some others.
Maybe we should lock all that down, but I don't see any percentage in
fixing just one place.
btw, I just noticed that chr() doesn't complain about arguments
exceeding 255 ...
regards, tom lane