[
https://issues.apache.org/jira/browse/LUCENE-3628?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13180294#comment-13180294
]
Uwe Schindler commented on LUCENE-3628:
---------------------------------------
I like the new code in MultiDocValues. The use of reflect.Array and
getComponentType/newInstance is fine here, as its only executed when you
actually request the merged array, which is expensive by default. If you access
DocValues through the standard API, not direct via array, all is as usual.
Merging the arrays is slow by default - so who cares? (<- this note is for
[~rcmuir]).
I think we should fix the DirectoryReader test to not use MultiNorms, maybe its
a specific test only for MultiNorms, so obsolete? I will look into it!
Additional note from discussion with [~simonw] on IRC: Maybe we should have the
same DV cache in SlowMultiReaderWrapper for non-norm DocValues, too.
> Cut Norms over to DocValues
> ---------------------------
>
> Key: LUCENE-3628
> URL: https://issues.apache.org/jira/browse/LUCENE-3628
> Project: Lucene - Java
> Issue Type: Improvement
> Components: core/index, core/search
> Affects Versions: 4.0
> Reporter: Simon Willnauer
> Assignee: Simon Willnauer
> Fix For: 4.0
>
> Attachments: LUCENE-3628.patch, LUCENE-3628.patch, LUCENE-3628.patch,
> LUCENE-3628.patch, LUCENE-3628.patch, LUCENE-3628.patch
>
>
> since IR is now fully R/O and norms are inside codecs we can cut over to use
> a IDV impl for writing norms. LUCENE-3606 has some
> [ideas|https://issues.apache.org/jira/browse/LUCENE-3606?focusedCommentId=13160559&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13160559]
> about how this could be implemented
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
