Hi Moritz, please use the latest pywps and wps-grass-bridge svn versions. Can you please verify if the mime type and the schema of the input vector are set in the WPS request generated by qgis? More below:
2011/4/15 Moritz Lennert <mlenn...@club.worldonline.be>: > On 15/04/11 10:59, Soeren Gebbert wrote: >> >> Hello Moritz, >> the problem is the temporal GML file generation in qgswpstools.py. >> This piece python code should solve the problem, which is also >> included in the patch: >> >> qgswpstools.py at line ~230 >> >> # Check in case a ".g" was attached to the tempfile >> if os.path.isfile(tmpFile + ".g") == True: >> tmpFile = tmpFile + ".g" > > Thanks for the patch ! I hope it works correctly. > > - Running the grass processes however gave me the same error message. > > This was using grass7 trunk and qgis trunk. > > - I've installed the Debian stable 1.4 version of QGIS. > > - I set up a simple pywps instance with only two of the example processes > from the source code: ultimatequestionprocess and buffer. Both run, but the > buffer process gives completely wrong buffers (much too large - maybe it is > working in degrees instead of meters, even though the QGIS project has units > meters and the correct projection defined). Unfortunately i am not familiar with the default processes of PyWPS. > > - Trying to run v.perturb or v.buffer without or with your patch applied > gives me > > <ows:Exception exceptionCode="NoApplicableCode"> > <ows:ExceptionText>Failed to execute WPS process > [v.perturb]: Unsupported MimeType: "text/xml". Unable to create input > location from input ./pywpsInput0DORwt</ows:ExceptionText> > </ows:Exception> This error is related to the wps-grass-bridge. It requires the definition of the mime type and the schema in case of XML data to identify GML or KML data. Because QGIS always produce GML this error should not occur in case the mime types are set correctly in the request. In case the request is correct the error is located in the wps-grass-bridge and i can fix it in the latest svn version. > > I'm currently recompiling a fresh a fresh checkout from qgis trunk in order > to test with that. Thanks for your effort and testing. The main problem is that all required software is in experimental state so the identification of the problem is quite hard. Best regards Soeren > > I'll keep you posted. > > Moritz > _______________________________________________ grass-dev mailing list grass-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-dev
