Gerrit Holl

Even easier test: checking overlap with itself also runs into an infinite loop: `AreaDefBoundary(germ).contour_poly.intersection(AreaDefBoundary(germ).contour_poly)`

It's supposed to be `(lon, lat)`. From https://pyresample.readthedocs.io/en/latest/geo_def.html: > It can be tested if a (lon, lat) point is inside a GeometryDefinition > > ``` > >>> print((0, -90) in...

Although that doc applies to `GeometryDefinition`, I assume that `AreaDefinition` uses the same units (the docstring for `AreaDefinition.__contains__` doesn't say).

Yes, it's full disk ABI, but I also encountered the problem with less-than-full-disk areas, probably whenever not all corners are valid lat/lons.

This is already supported with `area.crs.to_cf()`? Or is something missing? Apart from some bugs, it appears to work.

This also affects `__contains__`: ``` from pyresample.area_config import parse_area_file area = parse_area_file( "/data/gholl/checkouts/satpy/satpy/etc/areas.yaml", "seviri_0deg")[0] print((0, 0) in area) print(area.outer_boundary_corners) ``` fails already at the `(0, 0) in area` line: ```...

`ar.dump()` and `ar2.dump()` are actually equal: ``` test: description: test projection: proj: eqc lat_ts: 0 lat_0: 0 lon_0: 180 x_0: 0 y_0: 0 a: 6378136.6 rf: 298.257006177324 no_defs: null type:...

It seems information about the CRS has gone missing: ``` In [255]: ar.crs Out[255]: Name: Earth (2015) / Ographic / Equirectangular, clon = 180 Axis Info [cartesian]: - E[east]: Easting...

Is there an issue yet for changing `dump` to use WKT instead of a proj4 dict?

For me pyrasite-shell was hanging on Python 3.9, apparently due to missing gdb. Downgrading to Python 3.8 and `conda install gdb` resolved the problem (I could not `conda install gdb`...