[gdal-dev] Problem with gdal 1.8 -- segmentation fault on large image with gdaladdo

Stephen Mather mather.stephen at gmail.com
Sat Jan 1 11:34:57 EST 2011


Seems large enough for me.

Stephen Mather

On Jan 1, 2011, at 5:47 AM, Even Rouault <even.rouault at mines-paris.org> wrote:

>> 
>> What's the new ceiling on image size for overview calculation?
> 
> The same as before hopefully. I'm not sure what the limit is. From a quick look
> at the code, I'd say perhaps if the raster_width * 64 > 2 billion. This could be
> probably overcomeby modifying the algorithm if it became a limit in real use
> cases, but I'm not aware of someone having reported it up to here...
> 
>> 
>> Steve
>> 
>> 
>> 
>> On Fri, Dec 31, 2010 at 11:38 AM, Even Rouault <even.rouault at mines-paris.org
>>> wrote:
>> 
>>> Stephen,
>>> 
>>> this should be fixed now in SVN.
>>> 
>>> This was indeed something wrong in a recent optimization in overview
>>> computation. This could cause integer overflow, that was triggered when
>>> dealing
>>> with a dataset where width * height > 2 billion.
>>> 
>>> Thanks for reporting !
>>> 
>> 
> 
> 


More information about the gdal-dev mailing list