Обсуждение: Anyone else seeing CREATE TABLE crashes with psqlodbc?

I am seeing repeated seg faults when using psqlodbc to issue CREATE TABLE statements to postgresql.

My system is:

RedHat 9
Postgres 7.3
unixODBC 2.2.3
pgsqlODBC 7.2.5

The CREATE TABLE command is being executed, but the crash occurs as pgsqlODBC is attempting to
tear-down the statement (it appears).  The crash occurs variously inside the pgsqlODBC driver,
sometimes when it tries to free() in PG_SQLFreeStmt, QR_Destructor (qresult.c) or perform a strlen() in SQLRowCount(),
trim()(odbcapi.c).  One presumes in both cases invalid pointers are being passed around (a bad int* to free() and a bad
char*to trim()/strlen()). 

This is a major problem for us and could completely derail our plans to move to PostGres.

I am wondering if anyone else with a similar system is seeing this?

Thanks.

---
Clay
Cisco Systems, Inc.
claycle@cisco.com
(972) 813-5004


I've stopped 14,313 spam messages. You can too!
One month FREE spam protection at http://www.cloudmark.com/spamnetsig/}

** Reply to message from "Clay Luther" <claycle@cisco.com> on Mon, 28 Apr 2003
14:09:47 -0500

Hi Clay,
 If you haven't yet, try fooling around with some of the psqlodbc parameters.
 We were seeing a fair number of gpf's from our app running through the windows
client, untill we turned on the dissalow premature setting for the client. Most
gpfs were resulting from an attempt by our app to do a "free statement" on a
statement handle that hat already been closed by the driver.

Caveat on this is that I'm not sure I've been successful in geting the unixodbc
driver to listen to any parameters from the odbc.ini file at all :)

Regards,
Wayne


> I am seeing repeated seg faults when using psqlodbc to issue CREATE TABLE statements to postgresql.
>
> My system is:
>
> RedHat 9
> Postgres 7.3
> unixODBC 2.2.3
> pgsqlODBC 7.2.5
>
> The CREATE TABLE command is being executed, but the crash occurs as pgsqlODBC is attempting to
> tear-down the statement (it appears).  The crash occurs variously inside the pgsqlODBC driver,
> sometimes when it tries to free() in PG_SQLFreeStmt, QR_Destructor (qresult.c)
> or perform a strlen() in SQLRowCount(), trim() (odbcapi.c).  One presumes in
> both cases invalid pointers are being passed around (a bad int* to free() and a
> bad char* to trim()/strlen()).
>
> This is a major problem for us and could completely derail our plans to move to PostGres.
>
> I am wondering if anyone else with a similar system is seeing this?
>
> Thanks.
>
> ---
> Clay
> Cisco Systems, Inc.
> claycle@cisco.com
> (972) 813-5004
>
>
> I've stopped 14,313 spam messages. You can too!
> One month FREE spam protection at http://www.cloudmark.com/spamnetsig/}
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org

Clay Luther wrote:
>
> I am seeing repeated seg faults when using psqlodbc to issue CREATE TABLE statements to postgresql.
>
> My system is:
>
> RedHat 9
> Postgres 7.3
> unixODBC 2.2.3
> pgsqlODBC 7.2.5
>
> The CREATE TABLE command is being executed, but the crash occurs
> as pgsqlODBC is attempting to tear-down the statement (it appears).

Could you send me the Mylog output ?

regards,
Hiroshi Inoue
    http://www.geocities.jp/inocchichichi/psqlodbc/