alisterburt

Thanks again @mstabrin ! Will merge here when CI has finished

Following [discussion on zulip](https://napari.zulipchat.com/#narrow/stream/212875-general/topic/imshow.20return.20type/near/296525428) it seems the 'multiple dispatch' option has won out! @kcpevey it would be great to get this in over the next week so that it can...

@kcpevey based on what @brisvag says I think we can safely remove the commented line so have done - let me know if that's not the case @brisvag I think...

Hm, I don't quite see why you would expect the 3D data to give a 2D bounding box unless the image layer drops the dimensionality?

so you currently get a 3D bounding box for 3D data and a 2D one for 2D data? The `_display_*` methods slice the nD object (in this case bounding box)...

I see your point but idk, to me it would seem much weirder if the extents of the sliced data had different dimensionality than the data itself 🤷‍♂️ Maybe we...

Agreed @Czaki - I wasn't suggesting literally that we perform this calculation on the layer, rather that the slicing machinery could leave the necessary info somewhere accessible :)

looks awesome @perlman ! Does this mean we would potentially be able to leverage linear interpolation without messing up labels or not really?

Definitely don't break gray_r please! It's generally useful for visualising data with contrast based on absorption where the features of interest are dark-on-light :)

Thanks for the reviews @Czaki and @brisvag - I second @Czaki's suggestion for basic usable camps here followed by an issue/PR for discussion Cc @jni as I know he feels...