non-google bug report, postgres seg fault when comparing geometries

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

non-google bug report, postgres seg fault when comparing geometries

Cliff Wright
The bug tracker kept asking me to get a google account to report a bug,
which I do not want, so am using this process instead.
When using mapnik to display a NOAA nautical chart a postgres server seg faulted while comparing geometries from 2 different tables.
I did not isolate the exact comparison, however the core dump did indicate
which line and file caused the problem. looking at the code I see a memory
free was called apparently pointing at the wrong object. The same free
is used a couple of lines later pointing at the correct object. The patch
attached shows the fix I used. I was using version 1.3.1, however see the same
bug in subversion.



--
Cliff Wright <[hidden email]>

patcha.txt (568 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: non-google bug report, postgres seg fault when comparing geometries

Paul Ramsey
Thanks, filed. Should be in 1.3.3

On Fri, Mar 28, 2008 at 11:16 AM, Cliff Wright <[hidden email]> wrote:

> The bug tracker kept asking me to get a google account to report a bug,
>  which I do not want, so am using this process instead.
>  When using mapnik to display a NOAA nautical chart a postgres server seg faulted while comparing geometries from 2 different tables.
>  I did not isolate the exact comparison, however the core dump did indicate
>  which line and file caused the problem. looking at the code I see a memory
>  free was called apparently pointing at the wrong object. The same free
>  is used a couple of lines later pointing at the correct object. The patch
>  attached shows the fix I used. I was using version 1.3.1, however see the same
>  bug in subversion.
>
>
>
>  --
>  Cliff Wright <[hidden email]>
>
> _______________________________________________
>  postgis-devel mailing list
>  [hidden email]
>  http://postgis.refractions.net/mailman/listinfo/postgis-devel
>
>

Reply | Threaded
Open this post in threaded view
|

Re: non-google bug report, postgres seg fault when comparing geometries

Mark Cave-Ayland-4
> Thanks, filed. Should be in 1.3.3

Yes, that's fairly straightforward (I assume this is within the
LWGEOM_same function and not somewhere else? The diff doesn't quite match
up with SVN head and you didn't give a test query). Fix committed to SVN -
thanks for the bug report.


ATB,

Mark.

--
Mark Cave-Ayland
Sirius Corporation - The Open Source Experts
http://www.siriusit.co.uk
T: +44 870 608 0063


Reply | Threaded
Open this post in threaded view
|

Re: non-google bug report, postgres seg fault when comparing geometries

Cliff Wright
Yes, within LWGEOM_same is the place I ment.

On Fri, 28 Mar 2008 19:07:28 -0000 (GMT)
"Mark Cave-Ayland" <[hidden email]> wrote:

> I assume this is within the
> LWGEOM_same function and not somewhere else?