One more thought: We had some kind of FORCE 2D hint that WMS could use to
ask for the data back as 2D (ignoring the 2.5 D for speed).

Perhaps if you can figure out how to get oracle to respect that (with an
SQL function) we can get the curves back from the database as 2D...).  I
can find some examples <https://community.oracle.com/message/962492#962492>
online but it does not look fun :)

--
Jody Garnett

On 27 August 2015 at 18:27, Olle Markljung <[email protected]> wrote:

> It's not alot but it might happen and since everything then breaks the
> impact is quite profound.
> It's an Oracle NG datastore. When making offsets Oracle Spatial might
> create arcs.
>
> Before no one noticed since images from GeoServer would be without these
> and GML-responses would have no geometry node.
> Or perhaps someone noticed but accepted it due to the fact that arcs was
> not supported.
> But now nothing is rendered if one little feature has got at tiny little
> arc somewhere in the queried set.
>
> It would be more valuable if these where not skipped but that higher
> dimensions where ignored.
> Which I believe is the case if this message would be logged as a warning
> instead of an exception
> *"Circular strings are restricted to 2 dimensions at the moment.
> Contributions to get ND support welcomed!*"
>
> When we remove the exception locally the arc is rendered. That's an
> improvement since 10.2 from our perspective.
> I'll file a report tomorrow.
>
> Thanks for your attention!
>
> On Fri, Aug 28, 2015 at 2:51 AM, Jody Garnett <[email protected]>
> wrote:
>
>> Interesting, from my perspective this was a new feature and I did not
>> expect any regressions. The exception makes a bit of sense, unable to
>> interpolate the 3D positions.
>>
>> I am curious what store you are using for your data that has lots of
>> circular strings. How did it work previously that the geometry was just not
>> returned? Did you get null or something ... or did the feature reader skip
>> over them?
>>
>> I recommend making a bug report - and look at getting a pull request that
>> either does the 3D interpolation or "skips" over geometry that cannot be
>> decoded. This would be preferable to adding a bunch of exception handling
>> in your code.
>> --
>> Jody
>>
>> --
>> Jody Garnett
>>
>> On 27 August 2015 at 15:18, Olle Markljung <[email protected]> wrote:
>>
>>> Hello,
>>>
>>> I'm eager to find out why CircularString throws an exception instead of
>>> just logging a warning when confronted with geometries in 3D (or more).
>>>
>>> https://github.com/geotools/geotools/blob/master/modules/library/api/src/main/java/org/geotools/geometry/jts/CircularString.java#L90
>>>
>>> In 10.2 we didn't get any exception. The geometry was just not handled
>>> and returned as null thus not rendering in images and no geometry node in
>>> GML-output from GeoServer.
>>>
>>> Now we do and we have to handle this exception everywhere.
>>> No images at all from GeoServer and exception in GML-output if one
>>> feature with 3D CircularRing is in the query.
>>> If we remove the throwing of exception we get an image and the 3D
>>> geometry is returned in its 2D form. This is far more useful to us since it
>>> renders in images and gives us something that is almost correct in
>>> GML-output.
>>>
>>> Would it be OK to just log this as a warning instead of throwing an
>>> exception?
>>> Perhaps some kind of option otherwise?
>>>
>>> Regards,
>>> Olle Markljung
>>>
>>>
>>> ------------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> GeoTools-Devel mailing list
>>> [email protected]
>>> https://lists.sourceforge.net/lists/listinfo/geotools-devel
>>>
>>>
>>
>
------------------------------------------------------------------------------
_______________________________________________
GeoTools-Devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to