[PROJ] PROJ 7.0.0RC1

Even Rouault even.rouault at spatialys.com
Wed Feb 19 14:02:19 PST 2020


On mercredi 19 février 2020 08:11:55 CET Kristian Evers wrote:
> > On 18 Feb 2020, at 13:45, Even Rouault <even.rouault at spatialys.com> wrote:
> >> Hence it seems that proj-datumgrid itself is not entirely superseded by
> >> proj-data, unlike the regional archives.
> > 
> > For "normal" users, proj-datumgrid should be deprecated. We indeed need a
> > few "old" grids for the tests, that can be fetched for there. We could
> > potentially consider just dumping them, or a spatial subset of them, in
> > PROJ main repo, now that we have a data/for_tests subdirectory.
> 
> That is probably the simplest solution. Do you have clear image of which
> grids that would be?

From the checks in test/cli/Makefile.am:
- conus
- ntv1_can.dat
- MD
- ntf_r93.gsb
- egm96_15.gtx
- ntv2_0.gsb

The 2 later are a bit big though. We may need to extract a subset of them to 
the area(s) actually tested. A simple gdal_translate would do for 
egm96_15.gtx. Would be a bit more tricky with ntv2_0.gsb and its subgrids.
I can potentially have a try at that next week.

> Also, should these test files be added to the release
> tarballs?

Ideally yes, I'd say. That could simplify test/cli/Makefile.am logic to 
unconditionally run the tests.

-- 
Spatialys - Geospatial professional services
http://www.spatialys.com


More information about the PROJ mailing list