I believe this is a regression introduced by #217.
In rel8 1.4.0.0, calling nextval "foo.bar" to get the next value of a sequence bar in a schema foo used to generate the SQL nextval(CAST(E'foo.bar' AS text)), which is nextval('foo.bar').
However, in 1.4.1.0, it now generates nextval(CAST(quote_ident(CAST(E'foo.bar' AS text)) AS text)), which is nextval(quote_ident('foo.bar')) which is nextval('"foo.bar"'). This SQL is incorrect, and causes an error relation "foo.bar" does not exist because it's trying to access the sequence named "foo.bar" in the current schema, when in reality I want the sequence bar in schema foo.
I believe this is a regression introduced by #217.
In rel8 1.4.0.0, calling
nextval "foo.bar"to get the next value of a sequencebarin a schemafooused to generate the SQLnextval(CAST(E'foo.bar' AS text)), which isnextval('foo.bar').However, in 1.4.1.0, it now generates
nextval(CAST(quote_ident(CAST(E'foo.bar' AS text)) AS text)), which isnextval(quote_ident('foo.bar'))which isnextval('"foo.bar"'). This SQL is incorrect, and causes an errorrelation "foo.bar" does not existbecause it's trying to access the sequence named"foo.bar"in the current schema, when in reality I want the sequencebarin schemafoo.