Re: Leak repairs

> -----Original Message-----
> From: Anoop Kumar [mailto:anoopk(at)pervasive-postgres(dot)com]
> Sent: 19 July 2005 07:35
> To: Dave Page
> Cc: pgsql-odbc(at)postgresql(dot)org
> Subject: RE: [ODBC] Leak repairs
>
> Hi Dave,
>
> I have tested the patch from you, and it does make the perfmon graph
> less steep. I am applying the patch to CVS.

Great, thanks. Any joy tracking down the last little bit?

> Regarding the call to PGgetisnull: My mistake! I forgot to remove it,
> thanks for pointing it out.

No problem.

I think we should put out a second snapshot sometime, as soon as the
rest of this memory leak is fixed, and the Excel problem is sorted. Any
progress on that one?

Also we need to expose the libpq SSL option to the user, probably on the
main connection dialogue, and obviously as a connection string option.
At the same time, the protocol version options can be removed. I can
probably look at that if you are pushed for time - let me know what you
prefer.

Finally, the threading problem. After discussion with Bruce Momjian and
Magnus Hagander, we think that the --enable-thread-safety configure
option should be used on Windows as well as other platforms as it does
do more than just enable critical sections around non-thread safe API
calls. I am (as time allows!) working on fixing PostgreSQL to allow
this.

Regards, Dave

Have you seen this one?

1121712844: 17186: not freed: '0x404ef808|s1' (800 bytes) from
'qresult.c:421'
1121712844: 17186: not freed: '0x404f0c08|s1' (800 bytes) from
'qresult.c:421'
1121712844: 17186: not freed: '0x404fd008|s1' (3200 bytes) from
'qresult.c:421'
1121712844: 17186: not freed: '0x404fe008|s1' (3200 bytes) from
'qresult.c:421'
1121712844: 17186: not freed: '0x404ff008|s1' (3200 bytes) from
'qresult.c:421'
1121712844: 17186: not freed: '0x40500008|s1' (3200 bytes) from
'qresult.c:421'
1121712844: 17186: not freed: '0x40501008|s1' (3200 bytes) from
'qresult.c:421'

Line 421 allocates a buffer into QResultClass.backend_tuples.

backend_tuples will be set into NULL in QR_Destructor() with
"self->backend_tuples = NULL". I suspect, that the memory leak.

Marko Ristola

Dave Page wrote:

>
>
>
>
>>-----Original Message-----
>>From: Anoop Kumar [mailto:anoopk(at)pervasive-postgres(dot)com]
>>Sent: 19 July 2005 07:35
>>To: Dave Page
>>Cc: pgsql-odbc(at)postgresql(dot)org
>>Subject: RE: [ODBC] Leak repairs
>>
>>Hi Dave,
>>
>>I have tested the patch from you, and it does make the perfmon graph
>>less steep. I am applying the patch to CVS.
>>
>>
>
>
>Great, thanks. Any joy tracking down the last little bit?
>
>
>
>>Regarding the call to PGgetisnull: My mistake! I forgot to remove it,
>>thanks for pointing it out.
>>
>>
>
>No problem.
>
>I think we should put out a second snapshot sometime, as soon as the
>rest of this memory leak is fixed, and the Excel problem is sorted. Any
>progress on that one?
>
>Also we need to expose the libpq SSL option to the user, probably on the
>main connection dialogue, and obviously as a connection string option.
>At the same time, the protocol version options can be removed. I can
>probably look at that if you are pushed for time - let me know what you
>prefer.
>
>Finally, the threading problem. After discussion with Bruce Momjian and
>Magnus Hagander, we think that the --enable-thread-safety configure
>option should be used on Windows as well as other platforms as it does
>do more than just enable critical sections around non-thread safe API
>calls. I am (as time allows!) working on fixing PostgreSQL to allow
>this.
>
>Regards, Dave
>
>---------------------------(end of broadcast)---------------------------
>TIP 1: if posting/reading through Usenet, please send an appropriate
> subscribe-nomail command to majordomo(at)postgresql(dot)org so that your
> message can get through to the mailing list cleanly
>
>