I wrote:
> However ... in the wake of 618c16707, I wonder if we should consider
> an alternative definition, which is to NOT clear errorMessage when
> starting a new pipelined query. (That would be this patch minus
> the addition to pqPipelineProcessQueue.) Thanks to 618c16707,
> the returned error PGresult(s) should bear exactly the same contents
> either way. What would change is that PQerrorMessage(conn) would return
> the concatenation of all errors that have occurred during the current
> pipeline sequence. I think that's arguably a more useful definition
> than what we have now --- though obviously it'd require a docs change.
> It seems to fit with the spirit of the v14 changes to ensure that
> PQerrorMessage tells you about everything that happened during a
> failed connection attempt, not only the last thing.
After studying the pipeline mode some more, I no longer think that's
a great idea. ISTM we want pipeline mode to act as nearly as possible
the same as issuing the same series of queries non-pipelined. But
this redefinition would make it different. So, barring objection,
I'll push that cleanup patch as-posted.
regards, tom lane