distortion correction code zero pads around array edge instead of NaNs outside of FOV

[dsavransky]

Noticed this when looking at some datacubes - the distortion correction code appears to leave a partial border of pixels with value=zero around the outer edge of the array, which should be all NaNs since it's well outside the field of view. Probably does not matter in practice but since there are no DQ extensions on podc cubes (yet?) the distinction of valid vs. invalid lenslets is based on NaNs so strictly speaking these should be NaNs not zeros.

Attached image shows the partial border around the outside edge of the 281x281 array. But in practice this seems fairly negligible.

[dsavransky]

Recovering rm journals for issue:

kfollette: Since this is low priority, I'm setting the target version to 1.5