Need better mask validation for broadcastable arrays

[keflavich]

Broadcastable masks, e.g. 2D masks, are valid and should work, but right now the validation scheme stops that. I don't know how to do this appropriately yet, but I've asked on stackoverflow

[ChrisBeaumont]

I suggested one approach on SO, which lets you test if two shapes are broadcastable without allocating memory for an array to hold both shapes. I'm not sure if thats the kind of laziness you are after?

[keflavich]

Yes, it is. That's a great solution; I'll try to test it out. I'm surprised there's not something easier.

[ChrisBeaumont]

Me neither, I've looked for the same thing in the past

[keflavich]

@astrofrog I left this one off the list, but I think it's good to go too.

[astrofrog]

👍

[astrofrog]

@keflavich - you're going to kill me, but could you add changelog entries for all the PRs you merged today? :)