[Rd] Solved (RODBC): was Re: *** buffer overflow detected ***: /usr/lib64/R/bin/exec/R terminated on R 2.6.2 to 2.8.0: logging a bug?

[Prof Brian Ripley]

On Fri, 7 Nov 2008, Corrado wrote:

> Dear friends,
>
> problem solved after talking to the unixODBC maintainer.
>
> The problem is that the odbc driver (not unixODBC). The standard Mandriva
> installation uses the old default postgresql driver, which was embedded in
> the unixODBC, and not the newer odbc driver.
>
> As long as you update the driver (by downloading, building and installing
> manually) to the appropriate version:
>
> Here is the most important link: http://psqlodbc.projects.postgresql.org/
>
> Is it the case of putting note somewhere for other Mandriva users?

You mean as in the RODBC README file which says 'this driver seems 
unreliable (see the ChangeLog)'?

You *had* been warned, and no one who does not use Mandriva would have any 
idea that they were not using PostgreSQL's current driver (and that is the 
driver described in the RODBC README).

>
> On Friday 31 October 2008 19:13:18 Duncan Murdoch wrote:
>> On 10/31/2008 12:57 PM, Corrado wrote:
>>> On Friday 31 October 2008 16:29:13 you wrote:
>>>> See Prof. Ripley's message.  RODBC is passing a buffer and buffer size
>>>> to the ODBC driver, and that driver is overflowing the buffer.  You
>>>> could probably hide this bug by making the buffer larger, but the real
>>>> fix is for the driver to take account of the buffer size it is being
>>>> given.
>>>
>>> 1) How do you make the buffer bigger? Are we talking about increasing the
>>> -- --param=ssp-buffer-size=4 at compilation time? would that solve the
>>> problem, in your opinion?
>>
>> I don't know.  I wouldn't do that.
>>
>>> 2) Why is this problem RODBC specific, if the problem is in the odbc
>>> driver?
>>
>> Because the other packages pass bigger buffers, or generate different
>> messages?
>>
>> I
>>
>>> do agree that RODBC passes buffer / buffer size, but why do other
>>> applications work and RODBC does not? If the problem was with the driver,
>>> then every other application linking to the same database through the
>>> same driver would have the same problem, unless the other applications
>>> pass the call (+ buffer / buffer size) in a different way.
>>
>> Or if they ask the driver to do different things, or pass a different
>> size of buffer, etc.  The fact that this only occurs for RODBC suggests
>> there's a problem there, but the fact that the driver is using sprintf
>> on a limited size of buffer is far stronger evidence that the problem is
>> in the driver.
>>
>> Duncan Murdoch
>>
>>> Please, keep in mind I am just trying to solve a problem that other
>>> people may have, because I am using standard installation.
>>>
>>> Best,
>>> --
>>> Corrado Topi
>>>
>>> Global Climate Change & Biodiversity Indicators
>>> Area 18,Department of Biology
>>> University of York, York, YO10 5YW, UK
>>> Phone: + 44 (0) 1904 328645, E-mail: ct529 at york.ac.uk
>>>
>>> ______________________________________________
>>> R-devel at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-devel
>
>
>
> -- 
> Corrado Topi
>
> Global Climate Change & Biodiversity Indicators
> Area 18,Department of Biology
> University of York, York, YO10 5YW, UK
> Phone: + 44 (0) 1904 328645, E-mail: ct529 at york.ac.uk
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595