| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Peter Eisentraut <peter_e(at)gmx(dot)net> |
| Cc: | josh(at)agliodbs(dot)com, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: Test database for new installs? |
| Date: | 2004-11-19 15:57:22 |
| Message-ID: | 28519.1100879842@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Peter Eisentraut <peter_e(at)gmx(dot)net> writes:
> AFAICT, error codes are only accessible through PGresult. But if the
> connection attempt fails, you have at best a PGconn. This is the same kind
> of issue we have with frontends parsing the "no password supplied" message,
> because PQconnect cannot supply an error code in this case.
Oh, right. Yes, we should do something about that. I think we could
invent PQerrorField which is to PQerrorMessage as PQresultErrorField
is to PQresultErrorMessage. The reason I didn't do so during the 7.4
cycle is that errors generated internally within libpq wouldn't have any
field information, at least not without significant hacking on the libpq
sources.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Andreas Pflug | 2004-11-19 15:59:51 | Re: Test database for new installs? |
| Previous Message | Simon Riggs | 2004-11-19 15:45:40 | Re: Adding a suffix array index |