Stephen and Robert, while starting to work on https://jira.opendaylight.org/browse/INFRAUTILS-19, I've realized something I wanted to run by you / anyone else interested on list before I raise respective Gerrits:
In https://git.opendaylight.org/gerrit/#/c/65100/ we bumped it but it's already outdated, that's 3.2.3 but current is 3.2.5 already. But IMHO we should just remove the dependencyManagement for dropwizard.metric from odlparent, because: Today it's not much used anyway - only 2 artifacts in controller depend on it - /sal-clustering-commons & sal-distributed-datastore, and they could well just declare it themselves directly. The only other usage of io.dropwizard.metrics I've found in all of autorelease was in netvirt, but that one was fake / left-over - about to be gone with the wind via https://git.opendaylight.org/gerrit/#/c/65339/. Tomorrow, projects wishing to expose metrics will get use dropwizard.metrics by way of an (upcoming) infrautils.metrics dependency anyway, instead of a direct dependency to dropwizard.metrics in their POM. While we cannot enforce this of course, remove it from odlparent dependencyManagement would make it clearer IMHO. More importantly though, a quick test reveals that parent POM depenencyManagement appears to be "stronger" than an indirect transitive dependency - PITA. So, in this specific case, best IMHO jst not to have a parent POM depenencyManagement - it's not required, and will only cause problems to projects who are going to depend on infrautils.metrics. Makes sense to you & OK for you guys if I raise a controller change to add the fixed dependency, and another one to remove dropwizard.metric dependencyManagement from odlparent? Tx, M. -- Michael Vorburger, Red Hat vorbur...@redhat.com | IRC: vorburger @freenode | ~ = http://vorburger.ch
_______________________________________________ controller-dev mailing list controller-dev@lists.opendaylight.org https://lists.opendaylight.org/mailman/listinfo/controller-dev
