David Smiley

Yes, it's fairly recent that getContentStreams was deprecated in favor of `org.apache.solr.client.solrj.SolrRequest#getContentWriter`. But it'll be around thru the 7.x series for back-compat reasons. You could send the stream.body param as...

There are a *ton* of parameters for Solr and this is obscure. Are you sure we should load up the Admin UI for these? I think not.

Sounds good. It *is* document but it hasn't been released: https://github.com/apache/solr/blob/96c1577987412004b5df653bdebe2eb1d470e624/solr/solr-ref-guide/modules/query-guide/pages/highlighting.adoc#L119

> Once that is committed, we can then do what this PR does @HoustonPutman A change like that risks Haythem having to start over an additional time. I am encouraging...

@haythemkh I know you are using the Google Java Format IntelliJ plugin but not the IntelliJ code style that accompanies it, which ensures that the import wildcards are configured correctly...

PerReplicaStates is referenced in ClusterState, DocCollection, and Replica. Based on my very limited understanding of PRS, it's a part of the state of a collection no matter how it's retrieved...

A big obstacle I'm seeing is split packages between solrj & solrj-zookeeper with respect to javadocs. For example the javadocs for solr-core will now contain broken links to most solrj-zookeeper...

@sigram There is an abstraction `SolrCloudManager` and some related classes (DistributedQueue, DistribStateManager, NodeStateProvider) that you introduced once upon a time that don't seem to have anything to do with SolrJ...

For constants related to the collection's state, I put it in an inner interface DocCollection.CollectionSProps. The "S" there is for State. If that's removed, then it's possibly confused with CollectionProperties....

Other than a CHANGES.txt entry, I think this PR is ready.