aa_grid about arts HOT 4 CLOSED

@erikssonpatrick The convention that aa_grid must have a length of 1 for the 1-D case is not a bug, but has been there for a long time. There is an explicit check in DOAngularGridsSet This has been there at least since 2016 when Jana introduced that method.

from arts.

I agree that aa_grid=[] would be the correct choice for 1D, but only for the case of thermal radiation and DISORT. For DOIT you must have an aa_grid of size greater than 0 even for 1D atmosphere. Furthermore, if you have a direct source, aa_grid must be greater than 0 even for 1D atmospheres in Disort. But I can add a check in the DisortCalc-methods, that the size of aa_grid must be greater than 0.

from arts.

I agree with the idea of simpler interfaces, but I don't know what I can do about this. It seems like an internal Disort-interaction change. So I remove myself from this.

from arts.

Closing this for now. Needs to be revisited later with a larger scale refactoring of how all angular grids work.
Assert was fixed in #524.

from arts.