* Fix single {module} token rejection at nested depth (#20474)
A module type with a single {module} placeholder in component template
names could not be installed in a nested module bay (depth > 1) because
the form validation required an exact match between the token count and
the tree depth. This resolves the issue by treating a single {module}
token as a reference to the immediate parent bay's position, regardless
of nesting depth. Multi-token behavior is unchanged.
Refactors resolve_name() and resolve_label() into a shared
_resolve_module_placeholder() helper to eliminate duplication.
Fixes: #20474
* Address review feedback for PR #21740 (fixes#20474)
- Rebase on latest main to resolve merge conflicts
- Extract shared module bay traversal and {module} token resolution
into dcim/utils.py (get_module_bay_positions, resolve_module_placeholder)
- Update ModuleCommonForm, ModularComponentTemplateModel, and
ModuleBayTemplate to use shared utility functions
- Add {module} token validation to ModuleSerializer.validate() so the
API enforces the same rules as the UI form
- Remove duplicated _get_module_bay_tree (form) and _get_module_tree
(model) methods in favor of the shared routine
* Establish GFKSerializerField and replace get_* methods in circuits.py
* Set read_only=True
* Apply GFKSerializerField to all matching SerializerMethodFields
* Use GFKSerializerField for ObjectChangeSerializer.changed_object and EventRuleSerializer.action_object
* Create MACAddress model and migrations to convert existing .mac_address fields to standalone objects
* Add migrations
* All views/filtering working and documentation done; no unit tests yet
* Redo migrations following VLAN Translation
* Remove mac_address filter fields and add table columns for device/vm
* Remove unnecessary "bulk rename"
* Fix filterset tests for Device
* Fix filterset tests for Interface
* Fix tests on single-object forms
* Fix serializer tests
* Fix filterset tests for VMInterface
* Fix filterset tests for Device and VirtualMachine
* Move new field check into lookup_map iteration
* Fix general MACAddress filter tests
* Add GraphQL types/filters/schema
* Fix bulk edit/create tests (bulk editing Interfaces will be unsupported because of inheritance from ComponentBulkEditForm)
* Make mac_address read_only on InterfaceSerializer/VMInterfaceSerializer
* Undo unrelated work
* Cleanup unused IPAddress derived stuff
* API endpoints
* Add serializer objects to interface serializers
* Clean up unnecessary bulk create forms/views/routes
* Add SearchIndex and adjust indexable fields for Interface and VMInterface
* Reorganize MACAddress classes out of association with DeviceComponents
* Move MACAddressSerializer
* Enforce saving only a single is_primary MACAddress per interface/vminterface
* Perform is_primary validation on MACAddress model and just check if one already exists for the interface
* Remove form-level validation
* Fix check for current is_primary setting when reassigning
* Model cleanup
* Documentation notes and cleanup
* Simplify serializer and add ip_addresses
* Add to VMInterfaceSerializer too
* Style cleanup
* Standardize "MAC Address" instead of "MAC"
* Remove unused views
* Add is_primary field for bulk edit
* HTML cleanup and add copy-to-clipboard button
* Remove mac_address from Interface and VMInterface bulk-edit forms
* Add device and VM filtering
* Use combined assigned_object_parent in table to match structure of IPAddressTable
* Add GFK fields to MACAddressSerializer
* Reorganize "Addressing" sections to remove from proximity to "Device Components" and related groupings
* Clean up migrations
* Misc cleanup
* Add filterset test
* Remove mac_address field from interface forms
* Designate primary MAC address via a ForeignKey on the interface models
* Add serializer fields for primary_mac_address
* Update docs
---------
Co-authored-by: Jeremy Stretch <jstretch@netboxlabs.com>
