"Michael Paesold" <mpaesold@gmx.at> writes:
> I have come across a problem (bug?) with PL/pgSQL GET DIAGNOSTICS.
Hm. This seems to be SPI's version of the same definitional issue
we're contending with for status data returned from an interactive
query: SPI is currently set up to return the status of the last
querytree it executes, which is probably the wrong thing to do in the
presence of rule rewrites. But I'm hesitant to change SPI until we know
what we're going to do for interactive query status.
regards, tom lane