[ https://issues.jboss.org/browse/WFCORE-2317?page=com.atlassian.jira.plugi... ] Brian Stansberry commented on WFCORE-2317: ------------------------------------------ [~michpetrov] Nice! I haven't had time to look, but it's nice you are attacking this. Since you are working this, please assign it to yourself. "My assumption is that requires/alternatives is only meant for siblings (and as such list and map do not need to be validated)" Two answers -- the theoretical and the practical: Theoretical: The elements of the list and the values in the map should be validated individually for internal correctness, if those elements/values themselves use ObjectTypeAtributeDefinition. Those checks can be limited to cases where the enclosing list/map uses ObjectListAttributeDefinition or ObjectMapAttributeDefinition. There is no need to validate across list elements or map values. Practical: I see no uses of ObjectMapAttributeDefinition in WildFly. I'm glad, as maps of objects are fairly ugly. So dealing with it is not a priority. There are some uses of ObjectListAttributeDefinition. I don't see any uses of requires/alternatives in those, so dealing with lists is not a priority. I believe my IDE search would have caught any uses in the elytron subsystem, but it's good to double check, particularly for list. It's uses of list in the elytron subsystem that would provide some motivation to get the list handling in, since that's new API and correctness of security config is more important than most config. Otherwise I'd be fine with ignoring lists at this point. -- This message was sent by Atlassian JIRA (v7.2.3#72005)