[pgpool-hackers: 3236] Re: Imcompatible behavior with PostgreSQL

[Jesper Pedersen]

Hi,

On 2/5/19 10:02 PM, Tatsuo Ishii wrote:
> Currently Pgpool-II does not behave same as PostgreSQL if non existing
> prepared statement is given. As shown below, Pgpool-II sends fatal
> error message then terminates the connection to frontend, while
> PostgreSQL returns an error and continues the process. Do we want to
> fix this?
> 

I would say that pgpool should behave as PostgreSQL on the protocol 
layer, as some clients may actually depend on the behavior done by 
PostgreSQL. If pgpool is then put in-front the clients will stop working.

In this precise case I think PostgreSQL is correct (ERROR + ReadyForQuery).

Just my $0.02

Best regards,
  Jesper