Cache costly subqueries asynchronously [LUCENE-8213]

IndexOrDocValuesQuery allows to combine costly range queries with a selective lead iterator in an optimized way. However, the range query at some point gets cached by a querying thread in LRUQueryCache, which negates the optimization of IndexOrDocValuesQuery for that specific query.

It would be nice to see an asynchronous caching implementation in such cases, so that queries involving IndexOrDocValuesQuery would have consistent performance characteristics.

---
Migrated from [LUCENE-8213](https://issues.apache.org/jira/browse/LUCENE-8213) by Amir Hadadi, updated Nov 28 2019
Attachments: [0001-Reproduce-across-segment-caching-of-same-query.patch](https://apache.github.io/lucene-jira-archive/attachments/LUCENE-8213/0001-Reproduce-across-segment-caching-of-same-query.patch), [thetaphi_Lucene-Solr-master-Linux_24839.log.txt](https://apache.github.io/lucene-jira-archive/attachments/LUCENE-8213/thetaphi_Lucene-Solr-master-Linux_24839.log.txt)
Linked issues:
- #10045

Start with IndexOrDocValuesQuery and LRUQueryCache, then review the linked issue and the attached reproduction patch for the cross-segment caching behavior. The work is done when costly subqueries can be cached asynchronously without undermining IndexOrDocValuesQuery's optimized lead-iterator performance.

Written by the indexing model from the issue text.