Ian Barwick writes:
> That seems to imply though that this statement in the docs:
> "...where xxx is an octal number, is the character with the corresponding
> ASCII code."
> is not completely correct. Maybe it should read something like:
>
> "...where xxx is an octal number which corresponds to a valid character in the
> database encoding."
What it actually does is insert exactly one byte of the given value. The
problem is that you can easily construct invalid code sequences that way,
and I that should not be allowable.
What I would like better is if \xxx took xxx as a code point and converted
it to the appropriately encoded byte sequence. For single-byte encodings
that would mean no change, for Unicode it would make a lot more sense than
what it does now, but I don't know if that concept can be applied to all
character set encodings.
--
Peter Eisentraut peter_e@gmx.net