Looks good. Do we having hatching tests?

We do, but I don't think any of them check (e)ps output.

Does this still break #6228 ?

Does this still break #6228 ?

I don't think this breaks #6228.
The example program gives me the same results (as in the discussion) before and after this patch.
Though the results look strange.

)

Also, it would be better if someone else can confirm the results, since I only tried this on my own laptop.

Some clarification to this path from my side. When I plotted with hatching in PyCharm, its own viewer
hatched bar plots correctly. But when I opened the eps file directly on ubuntu (and windows too), or included in a Latex file, the hatches disappeared.

Some clarification to this path from my side. When I plotted with hatching in PyCharm, its own viewer
hatched bar plots correctly. But when I opened the eps file directly on ubuntu (and windows too), or included in a Latex file, the hatches disappeared.

I tried this myself using a newly installed Ubuntu 19.04 virtual machine.
The Document Viewer shows the eps correctly.

The EPS file is generated by MacOS (in the fish shell, rather than PyCharm) and transferred to the VM.
The script to generate EPS is as follows.

from matplotlib import pyplot as plt
from matplotlib.patches import Ellipse

fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ell = Ellipse((0.5, 0.5), 0.3, 0.4, 30, fill=False, hatch='/', facecolor='red', edgecolor='black')
ax.add_patch(ell)
fig.savefig('test.eps')

The EPS file is shown correctly in Win10 by an app called Pictures Opener.

So, would you please provide more details about your situation?