On 08/08/2019 16.38, Luca Ferrari wrote: > On Thu, Aug 8, 2019 at 4:25 PM Tom Lane <tgl@xxxxxxxxxxxxx> wrote: >> Given that both of these troublesome schemas have dashes in their >> names, the most obvious theory is that clone_schema() has some >> oversight about double-quoting schema names everywhere necessary. > According to this implementation > <https://www.postgresql.org/message-id/CANu8FiyJtt-0q%3DbkUxyra66tHi6FFzgU8TqVR2aahseCBDDntA%40mail.gmail.com> > the first query is performed with quote_ident and that's the initial > failure. However, the original post claimed it has been converted from > a function to a procedure, so it could be the quote_ident part has > been removed. > What is the source of clone_schema? The clone_schema is the one from the above quoted post, i.e. the last version by Melvin Davidson. Although it has been converted to a procedure, the code itself has not been changed. If you wish, I could attach it but the "$BODY$" block is identical to the original. > Luca > > > Thank you for your responses!
