Обсуждение: [PATCH] notice handler
This patch fill fix the notice handler so that the odbc driver doesn't output notices from the backend to stderr.
--- connection.c.old 2005-09-17 13:50:26.000000000 -0400
+++ connection.c 2005-09-17 13:40:11.000000000 -0400
@@ -2349,6 +2349,29 @@
}
#else
+
+static void
+CC_handle_notice(void *arg, const char *msg)
+{
+ QResultClass *qres;
+
+ qres = (QResultClass*)(arg);
+
+ if (qres == NULL)
+ {
+ // No query in progress, so just drop the notice
+ return;
+ }
+
+ if (QR_command_successful(qres))
+ {
+ QR_set_status(qres, PGRES_NONFATAL_ERROR);
+ QR_set_notice(qres, msg); /* will dup this string */
+ mylog("~~~ NOTICE: '%s'\n", msg);
+ qlog("NOTICE from backend during send_query: '%s'\n", msg);
+ }
+}
+
/*
* Connection class implementation using libpq.
* Memory Allocation for PGconn is handled by libpq.
@@ -3161,6 +3184,9 @@
}
/* free the conninfo structure */
free(conninfo);
+
+ /* setup the notice handler */
+ PQsetNoticeProcessor(self->pgconn, CC_handle_notice, NULL);
mylog("connection to the database succeeded.\n");
return 1;
}
@@ -3173,8 +3199,6 @@
PGresult *pgres;
char errbuffer[ERROR_MSG_LENGTH + 1];
int pos=0;
-
- pgres = PQexec(self->pgconn,query);
qres=QR_Constructor();
if(!qres)
@@ -3183,6 +3207,11 @@
QR_Destructor(qres);
return NULL;
}
+
+ PQsetNoticeProcessor(self->pgconn, CC_handle_notice, qres);
+ pgres = PQexec(self->pgconn,query);
+ PQsetNoticeProcessor(self->pgconn, CC_handle_notice, NULL);
+
qres->status = PQresultStatus(pgres);
/* Check the connection status */
@@ -3388,7 +3417,6 @@
}
-
#endif /* USE_LIBPQ */
"Scot Loach" <sloach@sandvine.com> writes:
> This patch fill fix the notice handler so that the odbc driver doesn't output notices from the backend to stderr.
Why is that a "fix" and not "breaking it"? It sounds pretty ill-advised
to me. In particular, notices sent when not inside a query are usually
pretty significant (ie, the backend telling you why it's about to cut
off your connection), and I can't see a justification for arbitrarily
suppressing them.
regards, tom lane
Thanks Scot, patch applied.
Regards, Dave
> -----Original Message-----
> From: pgsql-odbc-owner@postgresql.org
> [mailto:pgsql-odbc-owner@postgresql.org] On Behalf Of Scot Loach
> Sent: 17 September 2005 18:52
> To: pgsql-odbc@postgresql.org
> Subject: [ODBC] [PATCH] notice handler
>
> This patch fill fix the notice handler so that the odbc
> driver doesn't output notices from the backend to stderr.
>
> --- connection.c.old 2005-09-17 13:50:26.000000000 -0400
> +++ connection.c 2005-09-17 13:40:11.000000000 -0400
> @@ -2349,6 +2349,29 @@
> }
>
> #else
> +
> +static void
> +CC_handle_notice(void *arg, const char *msg)
> +{
> + QResultClass *qres;
> +
> + qres = (QResultClass*)(arg);
> +
> + if (qres == NULL)
> + {
> + // No query in progress, so just drop the notice
> + return;
> + }
> +
> + if (QR_command_successful(qres))
> + {
> + QR_set_status(qres, PGRES_NONFATAL_ERROR);
> + QR_set_notice(qres, msg); /* will dup
> this string */
> + mylog("~~~ NOTICE: '%s'\n", msg);
> + qlog("NOTICE from backend during send_query:
> '%s'\n", msg);
> + }
> +}
> +
> /*
> * Connection class implementation using libpq.
> * Memory Allocation for PGconn is handled by libpq.
> @@ -3161,6 +3184,9 @@
> }
> /* free the conninfo structure */
> free(conninfo);
> +
> + /* setup the notice handler */
> + PQsetNoticeProcessor(self->pgconn, CC_handle_notice, NULL);
> mylog("connection to the database succeeded.\n");
> return 1;
> }
> @@ -3173,8 +3199,6 @@
> PGresult *pgres;
> char errbuffer[ERROR_MSG_LENGTH + 1];
> int pos=0;
> -
> - pgres = PQexec(self->pgconn,query);
>
> qres=QR_Constructor();
> if(!qres)
> @@ -3183,6 +3207,11 @@
> QR_Destructor(qres);
> return NULL;
> }
> +
> + PQsetNoticeProcessor(self->pgconn, CC_handle_notice, qres);
> + pgres = PQexec(self->pgconn,query);
> + PQsetNoticeProcessor(self->pgconn, CC_handle_notice, NULL);
> +
> qres->status = PQresultStatus(pgres);
>
> /* Check the connection status */
> @@ -3388,7 +3417,6 @@
>
> }
>
> -
> #endif /* USE_LIBPQ */
>
>
> ---------------------------(end of
> broadcast)---------------------------
> TIP 1: if posting/reading through Usenet, please send an appropriate
> subscribe-nomail command to majordomo@postgresql.org
> so that your
> message can get through to the mailing list cleanly
>
> -----Original Message----- > From: pgsql-odbc-owner@postgresql.org > [mailto:pgsql-odbc-owner@postgresql.org] On Behalf Of Tom Lane > Sent: 17 September 2005 23:20 > To: Scot Loach > Cc: pgsql-odbc@postgresql.org > Subject: Re: [ODBC] [PATCH] notice handler > > "Scot Loach" <sloach@sandvine.com> writes: > > This patch fill fix the notice handler so that the odbc > driver doesn't output notices from the backend to stderr. > > Why is that a "fix" and not "breaking it"? It sounds pretty > ill-advised > to me. In particular, notices sent when not inside a query > are usually > pretty significant (ie, the backend telling you why it's about to cut > off your connection), and I can't see a justification for arbitrarily > suppressing them. It doesn't suppress them, it processes them as a non-fatal message so the app can deal with them rather than just dumping them to stderr. Regards, Dave.
"Dave Page" <dpage@vale-housing.co.uk> writes:
> It doesn't suppress them, it processes them as a non-fatal message so
> the app can deal with them rather than just dumping them to stderr.
So the comment stating that notices are dumped on the floor when there
is no query in progress is lying?
regards, tom lane
> -----Original Message----- > From: Tom Lane [mailto:tgl@sss.pgh.pa.us] > Sent: 17 September 2005 23:34 > To: Dave Page > Cc: Scot Loach; pgsql-odbc@postgresql.org > Subject: Re: [ODBC] [PATCH] notice handler > > "Dave Page" <dpage@vale-housing.co.uk> writes: > > It doesn't suppress them, it processes them as a non-fatal > message so > > the app can deal with them rather than just dumping them to stderr. > > So the comment stating that notices are dumped on the floor when there > is no query in progress is lying? <looks again> Ah, sorry. Missed that. Will sort it... /D
> -----Original Message----- > From: pgsql-odbc-owner@postgresql.org > [mailto:pgsql-odbc-owner@postgresql.org] On Behalf Of Dave Page > Sent: 17 September 2005 23:39 > To: Tom Lane > Cc: Scot Loach; pgsql-odbc@postgresql.org > Subject: Re: [ODBC] [PATCH] notice handler > > > > > -----Original Message----- > > From: Tom Lane [mailto:tgl@sss.pgh.pa.us] > > Sent: 17 September 2005 23:34 > > To: Dave Page > > Cc: Scot Loach; pgsql-odbc@postgresql.org > > Subject: Re: [ODBC] [PATCH] notice handler > > > > "Dave Page" <dpage@vale-housing.co.uk> writes: > > > It doesn't suppress them, it processes them as a non-fatal > > message so > > > the app can deal with them rather than just dumping them > to stderr. > > > > So the comment stating that notices are dumped on the floor > when there > > is no query in progress is lying? > > <looks again> Ah, sorry. Missed that. > > Will sort it... Done. Interestingly, the old version of the driver has always done the same thing. /D