"Daniel Verite" <daniel@manitou-mail.org> writes:
> This boils down to the fact that the current quote_ident gives:
> =# select quote_ident('test․table');
> quote_ident
> --------------
> "test․table"
> whereas the quote_ident patched as proposed gives:
> =# select quote_ident('test․table');
> quote_ident
> -------------
> test․table
> So this is what I don't feel good about.
This patch was originally proposed as a simple, cost-free change,
but it's becoming obvious that it is no such thing. I think
we should probably reject it and move on.
regards, tom lane