Juan Luis Cano Rodríguez

While investigating this issue I noticed another thing:  I get the same result regardless of the `masked` and `resampling`. Is this to be expected?

Also, `GeoRaster2.get_tile` (internally `GeoRaster2.get_window`) fails for in-memory rasters: ``` from rasterio.enums import Resampling import telluric as tl from telluric.constants import * rs1 = tl.GeoRaster2.open("../tests/data/raster/overlap1.tif", lazy_load=False) rs2 = tl.GeoRaster2.open("../tests/data/raster/overlap2.tif", lazy_load=False) rs2_alt...

Confirmed original issue: ``` rs2_4326.save("/tmp/rs2_4326.tif") tl.GeoRaster2.open("/tmp/rs2_4326.tif").get_tile(4377, 3038, 13) # Empty image ```

> I think you should open on each of the issues here an issue of its own Done: #118 and #119

(Notice that there is the combination of two problems here, see #23)

A good starting point for geodesic buffering is this exquisite paper "Algorithms for geodesics" by Charles F. F. Karney, where modern algorithms to compute the direct and inverse problems of...

Hit by this today, I might try and fix it.

There is a broader question here, that is: what to do with an operation that produces a result that is not representable in the original CRS? This is exactly the...

Cartopy already visualizes it correctly:  But the core problem of telluric is still that a `reproject` might turn a valid geometry into an invalid one: ``` In [6]: geod_buffer...

See also #154.