Could somebody clarify how the crs.db is maintained? I am a bit lost. There are: - scripts/qgis_srs.sh - it seems that it was used for initial import to resources/srs.db but it is not used anymore? - scripts/postprocess_srs.sql - it was probably called once after scripts/qgis_srs.sh but it is not used anymore? - resources/srs.db - occasionally updated add hoc by hand correcting wrong params or adding missing EPSG - src/crssync/main.cpp which calls QgsCoordinateReferenceSystem::syncDb() and updates resources/srs.db during compilation with latest CRS from GDAL
If I got it, the srs.db was created once by qgis_srs.sh and postprocess_srs.sql, now it is maintained by hand but also automatically updated (records are overwritten!?) by data from GDAL during compilation, right? If it really works like that, it seams to me that: - the result is a mix of CRS from various sources which are difficult/impossible to trace down - the changes applied to resources/srs.db may be overwritten by data from GDAL My question/proposal: should not it all work vice versa? I.e. fill an empty srs.db with the latest data from GDAL first and then to add the missing projections/params? All the additional projections added by QGIS should be maintained in a readable text file (not in an sqlite binary). Radim _______________________________________________ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer