RFC4: libtiff dependency required ?

Hi,

I know RFC4 is big creamy pie to digest. There's one discussion point I identified
at least, which is if we make libtiff a required dependency or just an optional one.

https://github.com/rouault/PROJ/blob/rfc4_remote_and_geotiff_grid/docs/source/community/rfc/rfc-4.rst#discussion-points

"""The advantage of making it required is that proj-datumgrid could only ship TIFF grids. If we
don't make it a requirement, we will have to manage .gtx / .gsb for the .zip / .tar.gz
delivered in the proj-datumgrid-XXXXX packages, and a .tif version hosted on the CDN.""""

From a pure PROJ developer point of view, having it required also simplifies the testing
story and maintenance burden.

Even if it is optional, I do think (hope!) that most binary distribution around would
build against it. Anyone building a C/C++ GIS stack has GDAL in it, and thus libtiff.
(The same holds for curl by the way.)

One thing I didn't mention is that support for reading .gtx/.gsb would remain, at least in a medium
term, as there are grids in those formats in the wild we don't (or can't) distribute. It is
just that we would only distribute .tif grids from now.

Small implementation detail: we would need to make sure that PROJ 6.3 database doesn't
point to the -latest.zip versions as it does currently - because after RFC4 would be done,
those would contain .TIF files -, but identify the last package versions that ship with .gtx/.gsb

Even

-- 
Spatialys - Geospatial professional services
http://www.spatialys.com
_______________________________________________
PROJ mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/proj