Composite Relationship Classes and Topology Issues

[VinceE]

I am confused by behavior I am experiencing when using a Composite Relationship Class (with forward messaging) in conjunction with a Topology.

Below I have a "parent" polygon (black) that has relationship classes with a "child" polygon (blue) and "child" point (black plus sign). Both are composite relationships, and both have FORWARD messaging turned on. Things work as expected, meaning when I rotate ONLY the parent polygon, the children rotate with it. There are no Feature Datasets, Subtypes, or Topologies involved in this example:

VinceE_0-1724965245256.png

The below example is more complex, and where I am having problems. There are subtypes and a topology involved.

There are again 3 feature classes:

1. wotus_poly [subtypes: WETLAND, WATER] - a "WETLAND" type shown in green
2. pool - no subtypes, shown in blue
3. test_plot_pt [subtypes: WETLAND, UPLAND] - two "WETLAND" and one "UPLAND" shown in black and red, respectively

Relationships (2):

wotus_poly one-to-many composite with pool, forward messaging turned on. A second relationship class with the same properties with the test_plot_pt.

VinceE_2-1724966384011.png   VinceE_3-1724966474182.png

Topology: also a topology rule set up that requires all WETLAND subtype test plots must be within a WETLAND subtype polygon.

VinceE_4-1724966594226.png

Here is what the structure in Catalog View looks like (this is within a Feature Dataset for the topology):

VinceE_5-1724966643094.png

Here is a visual of a possible arrangement, and one where the topology rule would be validated correctly:

VinceE_1-1724966033974.png

The issue is that the "FORWARD" messaging aspect of the relationships DOES NOT WORK as expected. Rotating the parent has no impact at all on the location/arrangement of the child pools or points. Is this because these features are involved in a Feature Dataset? Is it because there's a topology involved in at least two of these feature classes? Are subtypes problematic here?

VinceE_7-1724967358995.png

I have verified that DELETING the parent polygon, in green, will also delete the related children, as expected in a COMPOSITE relationship. So, there is some interference happening that prevents the geometric updates via forward messaging. Any help would be appreciated!