update the advection_fv4 solver to properly average initial conditions (#230)

this hooks the averaging into preevolve, just like we do with
compressible_fv4

Author: zingale

docs/source/advection-error.ipynb

@@ -62,6 +62,7 @@ new ideas.
    :caption: Examples
    :hidden:
 
+   advection-error.ipynb
    compressible-convergence.ipynb
 
 .. toctree::

docs/source/index.rst

@@ -0,0 +1 @@
+../advection/problems

pyro/advection_fv4/problems

@@ -58,3 +58,13 @@ def substep(self, myd):
             (F_y.v() - F_y.jp(1))/myg.dy
 
         return k
+
+    def preevolve(self):
+        """Since we are 4th order accurate we need to make sure that we
+        initialized with accurate zone-averages, so the preevolve for
+        this solver assumes that the initialization was done to
+        cell-centers and converts it to cell-averages."""
+
+        # we just initialized cell-centers, but we need to store averages
+        for var in self.cc_data.names:
+            self.cc_data.from_centers(var)