[Proj] Misplaced SHP obtained with PROJ 4.8.0 called from gvSIG CE

[Jorge Arevalo]

On Wed, Aug 21, 2013 at 8:15 PM, Hermann Peifer <peifer at gmx.eu> wrote:
> On 2013-08-21 17:04, Jorge Arevalo wrote:
>>
>> Hello,
>>
>> I'm trying to fix a bug related with PROJ in gvSIG CE. This desktop
>> client creates its own JNI wrapper, based on PROJ sources. The
>> reprojection of shapefiles using grid files works with the wrapper
>> based on PROJ 4.5.0, but it fails if we construct the same wrapper
>> based on any upper version of PROJ (PROJ 4.8.0, for example).
>>
>> Here, the use case that causes the problem:
>>
>> - We want to reproject this shapefile from epsg:23030 to epsg:25830:
>> https://dl.dropboxusercontent.com/u/6599273/gis_data/EPSG23030_data.zip
>>
>> - We want to use this grid file to perform the operation:
>> https://dl.dropboxusercontent.com/u/6599273/gis_data/sped2et.gsb
>>
>> The operation done with the JNI wrapper based on PROJ 4.5.0 works
>> fine. But if we use PROJ 4.8.0 as base, we get a misplaced file. Here,
>> a screenshot. In blue, the original shapefile. In green, the shapefile
>> correctly reprojected (PROJ 4.5.0). In red, the misplaced shapefile
>> (PROJ 4.8.0)
>>
>>
>> https://dl.dropboxusercontent.com/u/6599273/errors/gvsigce/gvsigce_misplacing_reproj.png
>>
>> Debugging, we found the problem is in pj_transform.c. I guess it's
>> related with the way we pass information to PROJ functions. I've made
>> a diff between both versions (pj_transform in PROJ 4.5.0 and PROJ
>> 4.8.0), but I see several changes, and I don't know where to focus.
>> Any clue?
>>
>> Many thanks in advance, and best regards,
>>
>
> I assume that this is your definition of the target SRS:
>
> +proj=utm +zone=30 +ellps=GRS80 +units=m +no_defs
>
> if yes, you'd have to change it to:
>
> +proj=utm +zone=30 +ellps=GRS80 +nadgrids=@null +units=m +no_defs
>
> Further reading: Why do I get different results with 4.5.0 and 4.6.0?
> http://trac.osgeo.org/proj/wiki/FAQ#WhydoIgetdifferentresultswith4.5.0and4.6.0
>
> Hermann
>
>

Hello Hermann,

Yes, I think that is definitely related with my problem. My SRS
definition string is incorrect. Now, I need to review the way it's
built in gvSIG CE. Many thanks for putting me in the right track!


-- 
Jorge Arévalo
http://geomati.co