Quick hack: for now, don't reexpand if the movement has been splitted.

This is a workaround, reposing on the fact that the split and defer solver set
good quantity on the simulation movement and that once a movement is splitted,
it's usually manipulated from solvers. Of course if parent movement (order or
packing list) is changed, there won't be any divergences at tax line level, but
as there will be a divergence at invoice line level, it's acceptable, anyway
it's better than the results we had before.


git-svn-id: https://svn.erp5.org/repos/public/erp5/trunk@26062 20353a03-c40f-0410-a6d1-a30d3c3de9de

product/ERP5/Document/TaxRule.py

@@ -101,7 +101,8 @@ class TaxRule(DeliveryRule):
         else:
           for existing_simulation_movement in \
                 existing_simulation_movement_list:
-            if existing_simulation_movement.getDelivery() is None:
+            if existing_simulation_movement.getDelivery() is None\
+                and not "split" in existing_simulation_movement.getId():
               existing_simulation_movement.edit(**property_dict)
 
     # Pass to base class