Move getting variant manager out of the loop

This mostly just makes it easier for me to debug where the manager is obtained from the deprecated singleton signal.

Contributes to issue CURA-6600.

plugins/XmlMaterialProfile/XmlMaterialProfile.py

@@ -608,6 +608,7 @@ class XmlMaterialProfile(InstanceContainer):
         # Map machine human-readable names to IDs
         product_id_map = self.getProductIdMap()
 
+        variant_manager = CuraApplication.getInstance().getVariantManager()
         machines = data.iterfind("./um:settings/um:machine", self.__namespaces)
         for machine in machines:
             machine_compatibility = common_compatibility
@@ -702,7 +703,6 @@ class XmlMaterialProfile(InstanceContainer):
                         if buildplate_id is None:
                             continue
 
-                        variant_manager = CuraApplication.getInstance().getVariantManager()
                         variant_node = variant_manager.getVariantNode(machine_id, buildplate_id,
                                                                       variant_type = VariantType.BUILD_PLATE)
                         if not variant_node:
@@ -725,7 +725,6 @@ class XmlMaterialProfile(InstanceContainer):
                         if hotend_name is None:
                             continue
 
-                        variant_manager = CuraApplication.getInstance().getVariantManager()
                         variant_node = variant_manager.getVariantNode(machine_id, hotend_name, VariantType.NOZZLE)
                         if not variant_node:
                             continue
@@ -777,7 +776,6 @@ class XmlMaterialProfile(InstanceContainer):
                             if buildplate_name is None:
                                 continue
 
-                            variant_manager = CuraApplication.getInstance().getVariantManager()
                             variant_node = variant_manager.getVariantNode(machine_id, buildplate_name, VariantType.BUILD_PLATE)
                             if not variant_node:
                                 continue