"Guillaume Smet" <guillaume.smet@gmail.com> writes:
> On 1/12/07, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> (2) there is already a generalized solution to this, it's called
>> log_min_error_statement.
> I didn't think of that when posting my message but Bruce seems to say
> that we can't use it in this case.
Dunno why he thinks that. But there is a point here that could use
improvement: shouldn't log_min_error_statement be measured on the same
scale as log_min_messages, ie, LOG is relatively high priority rather
than relatively low priority? As the code stands, you'd have to knock
it down to DEBUG1 in order to see the statement generating a LOG
message. This might be harmless (since messages below log_min_messages
won't generate log output at all), but it's surely a bit confusing.
regards, tom lane