test_polar_alignment fails. The tick 0.0 is now missing but I don't understand how the change is affecting it.

If you stick a print(fig.axes[0].yaxis.get_major_locator()) at the end of that test, before this PR it gives a FixedLocator, but after it gives a RadialLocator. Makes sense, but I guess that might be the root cause of the failing image test.

diff --git a/lib/matplotlib/projections/polar.py b/lib/matplotlib/projections/polar.py
index f20498b271..40f431aefb 100644
--- a/lib/matplotlib/projections/polar.py
+++ b/lib/matplotlib/projections/polar.py
@@ -444,7 +444,7 @@ class RadialLocator(mticker.Locator):
             if _is_full_circle_rad(*self._axes.viewLim.intervalx):
                 rorigin = self._axes.get_rorigin() * self._axes.get_rsign()
                 if self._axes.get_rmin() <= rorigin:
-                    return [tick for tick in self.base() if tick > rorigin]
+                    return [tick for tick in self.base() if tick >= rorigin]
         return self.base()
 
     def _zero_in_bounds(self):

☝️ fixes the test that's broken - seems like a sensible change to make, but I'm not sure if it breaks any other tests.

Thanks for figuring out. Actually, there is no tick at the center in all the polar examples.

From a quick look, I don’t fully understand the code you pointed to, but given this is under the is_full_circle_rad condition, it might be intentional to leave out the tick at the center. Will need to have a closer look at context conditions and history of that code.

cross check: MATLAB does not do this. So it’s at least not something built in to match their behavior.