2D Lazy masks fail when indexed in 1D

[keflavich]

This sort of masking + indexing:

med = cube.median(axis=0)
mask = cube > med

mcube = cube.with_mask(mask)

spec = mcube[:,1,1]

will result in a broadcast failure:

ValueError: operands could not be broadcast together with shapes (4,) (2,2)

because the comparison_value is being indexed with the same view as the data.

[keflavich]

Huh, no, it's not. This is a bug, but not for the reasons I imagined.

[coveralls]

Coverage increased (+0.06%) to 74.937% when pulling 31ce4ff on keflavich:issue392 into b0f2fe3 on radio-astro-tools:master.

[coveralls]

Coverage increased (+0.06%) to 74.937% when pulling 31ce4ff on keflavich:issue392 into b0f2fe3 on radio-astro-tools:master.