Hi Even, On Tue, Jun 2, 2020 at 2:43 PM Even Rouault <even.roua...@spatialys.com> wrote:
> Hi Sean, > > > > > An empty string is printed to the terminal after the second warning. Note > > > that behavior was different with rasterio 1.1.4, which > > > set CHECK_WITH_INVERT_PROJ=YES internally in an attempt to improve the > > > defaults. > > > > > > The warning comes from > > > > https://github.com/OSGeo/gdal/blob/2818ff0a49f65741a55b9cea79df496cc05f85e3/ > > > gdal/frmts/grib/gribdataset.cpp#L2149 . > > > > > > Is this sensitivity expected? > > > > I'd say this isn't completely unexpected if the corners of the dataset are > far from the projection center. There might be convergence issues in the > inverse projection method (or it might not try to converge down to the > precision we expect with CHECK_WITH_INVERT_PROJ=YES) > > > > CHECK_WITH_INVERT_PROJ is mostly a gdalwarp implementation detail to make > things work in some cases. I wouldn't advise to use it universally unless > you know you need to use it (one good reason is that not all PROJ > projection methods have an inverse method, even if most popular ones do > have one) > > > > Even > Got it. Thank you! -- Sean Gillies
_______________________________________________ gdal-dev mailing list gdal-dev@lists.osgeo.org https://lists.osgeo.org/mailman/listinfo/gdal-dev
