Re: buffer overflow in 08.04.0100 und 08.04.0200? - Mailing list pgsql-odbc

From Hiroshi Inoue
Subject Re: buffer overflow in 08.04.0100 und 08.04.0200?
Date
Msg-id 4B58FEA0.2040901@tpf.co.jp
Whole thread Raw
In response to buffer overflow in 08.04.0100 und 08.04.0200?  (Jan-Peter Seifert <Jan-Peter.Seifert@gmx.de>)
List pgsql-odbc
Hi,

Could you send me directly the Mylog output?

regards,
Hiroshi Inoue

Jan-Peter Seifert wrote:
> Hello,
>
> we ran into a problem with the new versions of psqlODBC (08.04.0100 and
> 08.04.0200), which causes a crash.
> For example one of our functions that dynamically generates queries does
> a 'RAISE NOTICE' of the query string before executing it. Sometimes this
> string gets very long and the driver crashes. In the CommLog the string
> is cut. The limit seems to be 4095. It somehow seems to be connected to
> the global socket size which obviously is 4096?
> However, it's strange that 8.3.400 is not crashing although the string
> seems to be truncated as well.
>
> We checked it with a function that is simply 'raise noticing' a string
> longer than 4100 characters/bytes.
>
> -------------------------------------
> CREATE OR REPLACE FUNCTION teststringlength() RETURNS CHAR AS
>
> $BODY$
>
> DECLARE
> BEGIN
>
>     RAISE NOTICE ' ... more than about 4100 characters ...';
>
> RETURN 'Success';
>
> END;
> $BODY$
>
> LANGUAGE 'plpgsql' VOLATILE;
> -----------------------------------------------
>
> It gets executed with 'SELECT teststringlength();' within a
> psqlODBC-connection that uses the same ODBC data source as the other
> psqlODBC-Version. In 08.04.0100 and 08.04.0200 it crashes ... in 8.3.400
> it doesn't.
> Is there maybe a new parameter that is missing, because we reused the
> old data source (I'll check tomorrow)?
>
> Thank you very much for any hints,
>
> Peter


pgsql-odbc by date:

Previous
From: the6campbells
Date:
Subject: Re: unixODBC, psqlodbcw.so, and Wine...
Next
From: Raiford@labware.com
Date:
Subject: Re: unixODBC, psqlodbcw.so, and Wine...