[Qgis-developer] QGIS Server WFS-T integration

[Marco Hugentobler]

Hi René-Luc

 >For the GML, I'll update QgsGeometry class ?

Yes, it will be good to have everything in one place (also for a 
potential redesign of QgsGeometry). But the GML methods should not be 
exported to the bindings, maybe you could add a comment in that 
direction to qgsgeometry.h.

Regards,
Marco

On 03.10.2012 10:03, rldhont wrote:
> Hi Marco,
>
> I updated my branch with the definition of QGS_NAMESPACE and 
> OGC_NAMESPACE.
>
> For the GML, I'll update QgsGeometry class ?
>
> For Query_String, I found the code with a google search.
>
> I test WFS-T with a jQuery XML POST and with QGIS.
>
> Cheers,
> René-Luc D'Hont
> 3Liz
>
> Le 03/10/2012 09:26, Marco Hugentobler a écrit :
>> Hi René-Luc
>>
>> Thank you for this nice addition, WFS-T will certainly be a great 
>> contribution to QGIS server!
>>
>> I tried to compile the wfs_server_transaction branch, but it fails 
>> with the following error message (QGS_NAMESPACE and OGC_NAMESPACE are 
>> not defined):
>>
>> /home/marco/src/Quantum-GIS/src/mapserver/qgswfsserver.cpp:718:35: 
>> Fehler: »QGS_NAMESPACE« wurde in diesem Gültigkeitsbereich nicht 
>> definiert
>> /home/marco/src/Quantum-GIS/src/mapserver/qgswfsserver.cpp:721:39: 
>> Fehler: »OGC_NAMESPACE« wurde in diesem Gültigkeitsbereich nicht 
>> definiert
>>
>>
>> Maybe there is a commit that you did not push to github yet?
>>
>> Then, looking at qgswfsserver.cpp, it seems that most of the 2000 
>> code lines deal with GML. It would be good to separate that into a 
>> GML handling class on its own (which could also be used from 
>> qgswfsprovider, at the moment, the gml reading seems to be duplicated).
>>
>> In QgsPostRequestHandler, the code looks for the environment variable 
>> QUERY_STRING. I thought QUERY_STRING is only defined in http get 
>> requests and in post requests, there is only CONTENT_LENGTH ?
>>
>> Which WFS client did you use for testing (QGIS desktop)?
>>
>> Regards,
>> Marco
>>
>>
>>
>>
>> On 01.10.2012 15:21, rldhont wrote:
>>> Hi all,
>>>
>>> I need some review for the WFS-T implementation in QGIS-Server.
>>> https://github.com/rldhont/Quantum-GIS/tree/wfs_server_transaction
>>> This branche adds the capability to the QGIS Server to understand 
>>> transaction, adds a way to define which WFS layer can be editable 
>>> and if the administrator authorize insert and delete, adds an 
>>> enhancements for the WFS provider to support WFS-T which needs extra 
>>> parameters.
>>>
>>> For the configuration on the WFS-T part, Update is possible if the 
>>> layer is publish as a WFS layer, Insert is possible only if Update 
>>> is possible and Delete is possible only if Insert is possible. You 
>>> can check Delete on a layer and it's automatically a WFS layer, or 
>>> uncheck Publied on a layer and all the checkbox are unchecked.
>>>
>>> I have testing QGIS-Server transaction with an ESRI ShapeFile and a 
>>> Spatialite database. The both doesn't return a good FID, but only 
>>> Spatialite has persistent FID. The Spatialite provider probably 
>>> needs an update.
>>> I test my QGIS-Server with QGIS and some JavaScript code.
>>>
>>> René-Luc D'Hont
>>> 3Liz
>>> _______________________________________________
>>> Qgis-developer mailing list
>>> Qgis-developer at lists.osgeo.org
>>> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>>
>> _______________________________________________
>> Qgis-developer mailing list
>> Qgis-developer at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>


-- 
Dr. Marco Hugentobler
Sourcepole -  Linux & Open Source Solutions
Weberstrasse 5, CH-8004 Zürich, Switzerland
marco.hugentobler at sourcepole.ch http://www.sourcepole.ch
Technical Advisor QGIS Project Steering Committee