[Bug]: quiver plotting behaves bizarrely with polar plots

Bug summary

There is no documentation, as far as I can tell, explaining how quiver should behave with polar plots, and using pyplot.quiver with a polar plot results in odd plotting behaviour.

Code for reproduction

import numpy as np
from matplotlib import pyplot as plt

fig, ax = plt.subplots(subplot_kw={"polar": True})

longitude = np.arange(5, 365, 10)
colatitude = 90 - np.ones_like(longitude) * 60

sample_x = np.ones_like(longitude)     # Northward
sample_y = np.ones_like(longitude)     # Eastward

q = ax.quiver(np.radians(longitude), colatitude, sample_x, sample_y, angles="xy")

Actual outcome

Expected outcome

I would expect the code to plot arrows which were vectors pointing all pointing northeast; i.e., I expect sample_x to be defined as the northward (poleward) vector and sample y to be defined as the eastward (azimuthal) vector.

Additional information

No response

Operating system

macOS Ventura 13.5

Matplotlib Version

3.7.1

Matplotlib Backend

Installed osx event loop hook. MacOSX

Python version

3.10.11

Jupyter version

4.0.1

Installation

conda

Comece executando a reprodução fornecida com matplotlib.pyplot.quiver em um Axes polar. Rastreie como as coordenadas polares e os vetores com angles="xy" são interpretados e, em seguida, verifique se as setas resultantes correspondem às componentes norte e leste solicitadas. Considera-se concluído quando o comportamento for corrigido ou claramente documentado, com uma verificação de regressão para este exemplo.