Re: [HACKERS] Improving overflow checks when adding tuple to PGresult Re: [GENERAL] Retrieving query results

[Michael Paquier]

On Wed, Aug 30, 2017 at 4:24 AM, Tom Lane  wrote:
> Michael Paquier  writes:
>> On Mon, Aug 28, 2017 at 3:05 PM, Michael Paquier
>>  wrote:
>>> Attached are two patches:
>>> 1) 0001 refactors the code around pqAddTuple to be able to handle
>>> error messages and assign them in PQsetvalue particularly.
>>> 2) 0002 adds sanity checks in pqAddTuple for overflows, maximizing the
>>> size of what is allocated to INT_MAX but now more.
>
> I've pushed these (as one commit) with some adjustments.

Thanks!

> Mainly,
> I changed PQsetvalue to report failure messages with PQinternalNotice,
> which is what already happens inside check_field_number() for the case
> of an out-of-range field number.  It's possible that storing the
> message into the PGresult in addition would be worth doing, but I'm
> unconvinced about that --- we certainly haven't had any field requests
> for it.

OK, fine for me.
-- 
Michael


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Re: [HACKERS] Improving overflow checks when adding tuple to PGresult Re: [GENERAL] Retrieving query results

[Tom Lane]

Michael Paquier  writes:
> On Mon, Aug 28, 2017 at 3:05 PM, Michael Paquier
>  wrote:
>> Attached are two patches:
>> 1) 0001 refactors the code around pqAddTuple to be able to handle
>> error messages and assign them in PQsetvalue particularly.
>> 2) 0002 adds sanity checks in pqAddTuple for overflows, maximizing the
>> size of what is allocated to INT_MAX but now more.

I've pushed these (as one commit) with some adjustments.  Mainly,
I changed PQsetvalue to report failure messages with PQinternalNotice,
which is what already happens inside check_field_number() for the case
of an out-of-range field number.  It's possible that storing the
message into the PGresult in addition would be worth doing, but I'm
unconvinced about that --- we certainly haven't had any field requests
for it.

regards, tom lane


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

[HACKERS] Improving overflow checks when adding tuple to PGresult Re: [GENERAL] Retrieving query results

[Michael Paquier]

On Mon, Aug 28, 2017 at 3:05 PM, Michael Paquier
 wrote:
> Attached are two patches:
> 1) 0001 refactors the code around pqAddTuple to be able to handle
> error messages and assign them in PQsetvalue particularly.
> 2) 0002 adds sanity checks in pqAddTuple for overflows, maximizing the
> size of what is allocated to INT_MAX but now more.
>
> pqRowProcessor() still has errmsgp, but it is never used on HEAD. At
> least with this set of patches it comes to be useful. We could rework
> check_field_number() to use as well an error message string, but I
> have left that out to keep things simple. Not sure if any complication
> is worth compared to just copying the error message in case of an
> unmatching column number.

As this change requires I think an extra lookup, I am moving the
discussion to -hackers with a proper subject and the set of patches
attached (and the test program). This patch is registered in the next
commit fest.
-- 
Michael


0001-Refactor-error-message-handling-in-pqAddTuple.patch
Description: Binary data


0002-Improve-overflow-checks-of-pqAddTuple-in-libpq.patch
Description: Binary data
/*
 * Script to test PQcopyResult and subsequently PQsetvalue.
 * Compile with for example:
 * gcc -lpq -g -o pg_copy_res pg_copy_res.c
 */

#include 
#include 
#include "libpq-fe.h"

#define DEFAULT_PORT	"5432"
#define DEFAULT_HOST	"/tmp"
#define DEFAULT_DB		"postgres"

int
main()
{
	char *port = getenv("PGPORT");
	char *host = getenv("PGHOST");
	char *dbname = getenv("PGDATABASE");
	char connstr[512];
	PGconn *conn;
	PGresult *res, *res_copy;

	if (port == NULL)
		port = DEFAULT_PORT;
	if (host == NULL)
		host = DEFAULT_HOST;
	if (dbname == NULL)
		dbname = DEFAULT_DB;

	snprintf(connstr, sizeof(connstr), "port=%s host=%s dbname=%s",
			 port, host, dbname);

	conn = PQconnectdb(connstr);

	if (PQstatus(conn) != CONNECTION_OK)
	{
		fprintf(stderr, "Connection to database failed: %s",
PQerrorMessage(conn));
		return 1;
	}

	res = PQexec(conn, "SELECT 1");

	/* Copy the resuld wanted, who care what that is... */
	res_copy = PQcopyResult(res, PG_COPYRES_TUPLES | PG_COPYRES_ATTRS);

	PQclear(res);
	PQclear(res_copy);

	PQfinish(conn);
	return 0;
}

-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers