Trac #19384: Type B rigged configuration bijection doesn't work when …

…spinor is not the only leftmost factor

The following breaks:
{{{
sage: RC = RiggedConfigurations(['B',3,1], [[3,1],[3,1]])
sage: RC._test_bijection()
sage: RC = RiggedConfigurations(['B',3,1], [[1,1],[3,1],[1,1]])
sage: RC._test_bijection()
}}}
This is because we don't keep the vacancy numbers when converting back
from the A,,2n-1,,^(2)^ rigged configurations and instead recompute them
based upon the original dimensions (because we abuse the elements and
work with a mutating copy of them to avoid creating numerous temporary
parents).

URL: http://trac.sagemath.org/19384
Reported by: tscrim
Ticket author(s): Travis Scrimshaw
Reviewer(s): Ben Salisbury

src/sage/combinat/rigged_configurations/bij_type_B.py

@@ -78,6 +78,15 @@ def run(self, verbose=False):
             <BLANKLINE>
             0[]0
             <BLANKLINE>
+
+        TESTS:
+
+        Check that :trac:`19384` is fixed::
+
+            sage: RC = RiggedConfigurations(['B',3,1], [[3,1],[3,1]])
+            sage: RC._test_bijection()
+            sage: RC = RiggedConfigurations(['B',3,1], [[1,1],[3,1],[1,1]])
+            sage: RC._test_bijection()
         """
         if verbose:
             from sage.combinat.rigged_configurations.tensor_product_kr_tableaux_element \
@@ -180,7 +189,7 @@ def run(self, verbose=False):
                         bij.ret_rig_con[i]._list[j] //= 2
                         bij.ret_rig_con[i].rigging[j] //= 2
                         bij.ret_rig_con[i].vacancy_numbers[j] //= 2
-                self.ret_rig_con = self.tp_krt.parent().rigged_configurations()(*bij.ret_rig_con)
+                self.ret_rig_con = self.tp_krt.parent().rigged_configurations()(*bij.ret_rig_con, use_vacancy_numbers=True)
                 # Make it mutable so we don't have to keep making copies, at the
                 #   end of the bijection, we will make it immutable again
                 self.ret_rig_con._set_mutable()