Tom Lane wrote:
> Neil Conway <neilc@samurai.com> writes:
> >> There may be existing clients out there that miscompute the password
> >> packet length. Right now that does no harm. With an Assert in place
> >> in the backend, it will cause a database system restart.
>
> > Good point. However, I still think a sanity check would be appropriate
> > here. How about an elog(WARNING) ?
>
> I think that elog(LOG) is probably the right thing. IIRC, at that point
> in startup we will not send anything short of ERROR to the client, so
> elog(WARNING) is pointless from the client's point of view --- and a LOG
> is actually more likely to get into the server's log than a WARNING.
Agreed. I did not add any of that code. The actual place you were
testing was not the startup packet but the password packet.
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073