Crosshair indistinguishable from gray background

[bnehoran]

It seems that the color of the cursor is determined by inverting the color of the image behind it.

While this works in many cases, whenever the image is gray, which happens when there is no data, the inverse of this gray is itself, and the cursor just disappears.

This is frustrating, because even in locations where there is no data, I still want to know where I'm pointing, for a number of uses.

A fix strategy would involve finding a color-to-color function that has no fixed points, and hopefully one that maximizes contrast.

Steps to reproduce:
Open a dataset to a location where it doesn't have data.
For example: https://goo.gl/UDxDRb
Move the cursor over the window. The cursor becomes "invisible".

Browser: Chrome 70
OS: tested on Windows 10

Edit: This issue also makes it very difficult to see the cursor on normal tissue at high mip levels (where the image tends to average out to look nearly gray)

[nkemnitz]

Hm... the cursor color and behavior might be an OS thing - Neuroglancer just uses the default crosshair cursor.
It's definitely easier to change the default background color: You can do that by adding "crossSectionBackgroundColor": "#606060" to the top-level JSON state object. (Or whatever color you prefer)

[nkemnitz]

Looks like the only option is a custom cursor. To make it work on all sorts of backgrounds, we could
❤️ : create an outlined version - which could still become nearly invisible if the underlying EM image closely resembles the
🎉 : let the user control the color - which means another setting that needs to be saved, and which the user has to be aware of (also need to check if CSS color blending even works with cursors)