[jboss-jira] [JBoss JIRA] (WFLY-10577) Undertow subsystem should not depend on clustering code.

[ https://issues.jboss.org/browse/WFLY-10577?page=com.atlassian.jira.plugin... ] Jeff Mesnil commented on WFLY-10577: ------------------------------------ [~pferraro] sorry, I was not precise enough. WFLY-10352 did not introduce the dependency from undertow to clustering.common, I just noticed it was there while I was looking at the new MSC Service API. I can understand the dependency to org.wildfly.clustering.service for web cache thingie. However org.jboss.as.clustering.common does not deal about that. It is an abstraction on top of the Core API, right? My concern is having subsystem code depends on another one for core features. That makes the whole codebase harder to maintain as we have different ways to do the same thing (dealing with MSC Service in that case). If the Core API (or the MSC Service API) does not provide the correct abstraction (or need another level of abstraction), these features should be done in a place that benefits all subsystems. By the way, the dependency from wildfly-undertow to wildfly-clustering-common is not direct and goes through wildfly-security > wildfly-clustering-infinispan-spi: {code} ~/Developer/wildfly/undertow (master)$ mvn dependency:tree | ack -B 5 clustering.common [INFO] | +- org.jboss.threads:jboss-threads:jar:2.3.2.Final:compile [INFO] | \- org.jboss:jboss-vfs:jar:3.2.12.Final:compile [INFO] +- org.wildfly:wildfly-security:jar:14.0.0.Beta1-SNAPSHOT:compile [INFO] | +- org.wildfly.core:wildfly-domain-management:jar:6.0.0.Alpha1:compile [INFO] | +- org.wildfly:wildfly-clustering-infinispan-spi:jar:14.0.0.Beta1-SNAPSHOT:compile [INFO] | | +- org.wildfly:wildfly-clustering-common:jar:14.0.0.Beta1-SNAPSHOT:compile ~/Developer/wildfly/undertow (master)$ {code} That's why I was not sure if the dependency was intentional or just a byproduct of being available... -- This message was sent by Atlassian JIRA (v7.5.0#75005)