Lorenzo Gaifas

> it seems to be a failure to triangulate some points that are randomly generated Ah, this is a known old issue... Well, we don't need to solve it here...

Merged! Thanks a lot @aganders3, also for upstreaming what possible to vispy :)

For context: this is *technically* not the *correct* fix, because it simply assumes that anything that results as a completely transparent label should be ditched (even non-background ones)... Now, why...

what confuses me is that I tried a whole bunch of changes to basically get to the pain where the *correct* value is set (instead of being a constant) to...

> could you share the code to print such values in opengl? Sorry, maybe I was unclear: I'm printing the value *just before* setting (in python). So yeah, I'm not...

I'm fairly sure this is handled by vispy... EDIT: yes, and they are all float32 in the first place :/

Ok, narrowing it down: the setting is working correctly; the values that appear to be incorrect is the actual value from sampling the texture. The reason `floatEqual` fails is because...

So basically, in my test code, the data should contain the following mapped values: ```py [0.125 0.625 0.875] ``` With mappings being: ```py In [53]: layer._label_color_index Out[53]: {1: 0.875, 2:...

> for 0 the default color is transparent and for none the default color is black. So objects that does not have set color are painted with black. Oh, I...

Interpolation is `nearest`, so no. The jagged edges you see are not due to interpolation, but to the size of the ray casting step being relatively big compared to the...