Artifact with mark_area() when coloring intervals

I am attempting to shade an area under a curve between two points to demonstrate an integral, but I get a strange plotting artifact:

import numpy as np
import altair as alt
import pandas as pd

x = np.arange(-1,1,0.01)
y = np.exp(x)
A = (-0.8<=x) & (x<0.8)

data = pd.DataFrame({'x':x, 'y':y, 'A':A})
alt.Chart(data).mark_area().encode(x='x', y='y', color='A:N')

The behavior is different if I set different limits for the shaded area:

A = (-1<=x) & (x<0)

data = pd.DataFrame({'x':x, 'y':y, 'A':A})
alt.Chart(data).mark_area().encode(x='x', y='y', color='A:N')

Perhaps I am misusing mark_area(), but I can’t find anything in the documentation or other issues discussing how to do this correctly, and the fact that the behavior is different in these two cases makes me think it’s a bug, not misuse.

[jakevdp]

The best way to avoid a space between the sections would probably be to use a layer with the filtered dataset:

import numpy as np
import altair as alt
import pandas as pd

x = np.arange(-1,1,0.01)
y = np.exp(x)
A = (-0.8<=x) & (x<0.8)

data = pd.DataFrame({'x':x, 'y':y, 'A':A})
base = alt.Chart(data).mark_area().encode(x='x', y='y')
base + base.transform_filter('datum.A').mark_area(color='darkorange')

[jakevdp]

Oh, your working example has space between the sections, so I thought that was what you were expecting.