[
https://issues.apache.org/jira/browse/SOLR-5780?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13916643#comment-13916643
]
Furkan KAMACI commented on SOLR-5780:
-------------------------------------
I've created a patch file for approach of "keeping up to date
concurrentlinkedhashmap" and run full tests whether it has a side effect or not.
> Solr should benefit from Guava 16.0.1
> -------------------------------------
>
> Key: SOLR-5780
> URL: https://issues.apache.org/jira/browse/SOLR-5780
> Project: Solr
> Issue Type: Improvement
> Components: SolrCloud
> Affects Versions: 4.7
> Environment: All.
> Reporter: Guido Medina
> Attachments: SOLR-5780.patch
>
>
> Solr is using concurrentlinkedhashmap v1.2 and Guava 14.0.1 at the same time,
> according to concurrentlinkedhashmap author(s), that project main objective
> is to introduce ideas and then when proven they are ported to Guava.
> concurrentlinkedhashmap v1.2 was designed for Java 5 and v1.4 for Java 6+
> which is the target version Solr 4.x requires, v1.4 had a great improvement
> in performance and memory impact compared to v1.2 which was ported to Guava
> (I strongly believe v16.0.1+ will do)
> *Pertinent material:*
> *
> [http://stackoverflow.com/questions/15299554/what-does-it-mean-that-concurrentlinkedhashmap-has-been-integrated-into-guava]
> * [https://code.google.com/p/concurrentlinkedhashmap/wiki/Changelog]
> All that said, concurrentlinkedhashmap should be eliminated _- OR keep up to
> date becauseit is the core of in-memory cache, same as Guava -_ and code
> using it should instead use MapMaker builder from Guava.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
