[postgis-users] Infinite loop in st_intersects - because of incorrect data out of st_transform?

[Magnus Hagander]

On Thu, Jun 9, 2011 at 12:35, Sandro Santilli <strk at keybit.net> wrote:
> On Wed, Jun 8, 2011 at 8:58 PM, Magnus Hagander <magnus at hagander.net> wrote:
>> On Tue, May 31, 2011 at 14:06, Magnus Hagander <magnus at hagander.net> wrote:
>>> On Fri, May 27, 2011 at 21:07, Sandro Santilli <strk at keybit.net> wrote:
>>>> On Fri, May 27, 2011 at 09:28:57AM -0700, Paul Ramsey wrote:
>>>>> I'm finding it difficult to have the regression test and also the
>>>>> NOTICE in isvalid and also have the differential behaviour for
>>>>> different versions of GEOS. One is going to have to go.  I can't
>>>>> ensure the "reason" I provide in my short circuit to match the reason
>>>>> GEOS is going to give.
>>>>
>>>> What you can do is _not_ using tickets.sql but a dedicated file, maybe
>>>> bugXXX.sql and bugXXX_pre_geos33.sql, and switch between the two from
>>>> the Makefile.
>>>
>>> So to be clear on this for someone who's not following the svn repo -
>>> has a fix for this been applied now?
>>
>> Hrrm... Ping?
>
> The fix is in upgrading GEOS (3.3.0 is out).
> If you can upgrade PostGIS you can also upgrade GEOS, right ?

Uh, that sounds like a major version rather than a patch version? Or
do I not know how GEOS versions works?

Applying a patch version is pretty trivial and makes no difference if
it's GEOS or PostGIS. But if it's a new major version, that would
require a whole new RPM-set to be released etc, and IIRC the version
of geos is more or less "locked" to the version of postgis at release
time? Not entirely sure though - Devrim, can you comment on that? GEOS
3.3.0 for older versions of postgis in the rpm repo, since it has an
important bugfix?


-- 
 Magnus Hagander
 Me: http://www.hagander.net/
 Work: http://www.redpill-linpro.com/