fix period-out; default to None for compat with categorical/sparse.

Author: ResidentMario

Diff for: pandas/core/generic.py

@@ -3373,7 +3373,7 @@ def convert_objects(self, convert_dates=True, convert_numeric=False,
 
     @Appender(_shared_docs['fillna'] % _shared_doc_kwargs)
     def fillna(self, value=None, method=None, axis=None, inplace=False,
-               limit=None, downcast=None, errors='coerce'):
+               limit=None, downcast=None, errors=None):
         inplace = validate_bool_kwarg(inplace, 'inplace')
 
         if not missing.maybe_fill(self, value, errors):

Diff for: pandas/core/missing.py

@@ -626,11 +626,10 @@ def maybe_fill(obj, value, errors):
     """
     from pandas import Series
     if isinstance(obj, Series):
-        if errors == 'coerce':
+        if errors is None or errors == 'coerce':
             return True
         else:
-            is_period = is_period_arraylike(obj)
-            if not valid_fill_value(value, obj.dtype, is_period=is_period):
+            if not valid_fill_value(value, obj.dtype):
                 if errors == 'raise':
                     raise TypeError('"value" parameter must be compatible '
                                     'with the {0} dtype, but you passed a '