[
https://issues.apache.org/jira/browse/SOLR-16677?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17708324#comment-17708324
]
Vinayak Hegde commented on SOLR-16677:
--------------------------------------
I'm interested in working on this, but I have some doubts. In Solr, we used to
override the IndexReader.document() and IndexReader.getTermVector() methods,
but they're now deprecated. What should we do in this case? Should I:
1. Add the @Deprecated annotation and leave the methods as it is. use the
alternative method for them (like IndexReader.storedFields().document() for
IndexReader.document()) Or
2. throw UnsupportedOperationException() from those methods and use the
alternative method for them
3. Is there another solution?
I'd appreciate your guidance on how to proceed.
> Update Solr to use new Lucene 9.5 storedFields() API
> ----------------------------------------------------
>
> Key: SOLR-16677
> URL: https://issues.apache.org/jira/browse/SOLR-16677
> Project: Solr
> Issue Type: Improvement
> Reporter: Michael Gibney
> Priority: Major
>
> Solr should migrate Lucene's {{LeafReader.storedFields()}} API ([new as of
> Lucene 9.5|https://github.com/apache/lucene/pull/11998]) in place of the
> now-deprecated {{LeafReader.document()}} API (and associated
> {{CloseableThreadLocal}} approach).
> A significant amount of heap space can be retained by the deprecated
> approach, particularly in many-segment, many-thread use cases (look for
> CompressedStoredFieldsReader in heap dumps).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
