On vendredi 8 mars 2019 18:22:42 CET Markus Metz wrote: > On Fri, Mar 8, 2019 at 5:56 PM Even Rouault <even.roua...@spatialys.com> > > wrote: > > On vendredi 8 mars 2019 17:01:57 CET Markus Metz wrote: > > > My question is, what should GRASS do with a SRS without proj string? The > > > best answer is probably: use WKT instead of a proj string. But most of > > the > > > > GRASS code uses (for historical reasons) some form of proj syntax, also > > to > > > > store projection info for a GRASS location. If there is no proj string, > > > GRASS thinks there is no SRS and falls back to a generic xy system > > which is > > > > not always correct. > > > > The WKT will not help here. You will be able to get here, but you won't be > > able to do coordinate transformation with it. This is just that this EPSG > > method is not implemented or mapped to a PROJ projection method > > I understand that coordinate transformation with PROJ will not be possible. > But having a valid SRS definition would at least allow to compare two > spatial references, i.e. OSRIsSame() should work. Having a valid SRS > instead of dumping it would also help if the corresponding PROJ method > becomes implemented in the future. That's why I am wondering about an > alternative to the current proj-like projection information in GRASS.
Yes, for having a reference SRS definition, you can use the WKT export. By "useful", I meant being able to do coordinate transformation. To compare SRS for equality, rather than using full string comparison, I'd suggest using proj_is_equivalent_to() which has a few options for different levels of equality. Even -- Spatialys - Geospatial professional services http://www.spatialys.com _______________________________________________ grass-dev mailing list grass-dev@lists.osgeo.org https://lists.osgeo.org/mailman/listinfo/grass-dev