[jira] [Commented] (LUCENE-5694) Fix Disjunction*Scorer not to score unless you ask it to

["Robert Muir (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/LUCENE-5694?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14005889#comment-14005889 ] 

Robert Muir commented on LUCENE-5694:
-------------------------------------

If you disable BS1 in luceneutil, it will appear this patch gives a speedup, but from my investigation this is just trickery.

I did further benchmarking to hone in and its just the same speed. 

But it can be beneficial in cases where score() is not called (ConstantScoreQuery, constructing a filter, when there are multiple prohibited clauses, etc etc)

> Fix Disjunction*Scorer not to score unless you ask it to
> --------------------------------------------------------
>
>                 Key: LUCENE-5694
>                 URL: https://issues.apache.org/jira/browse/LUCENE-5694
>             Project: Lucene - Core
>          Issue Type: Bug
>            Reporter: Robert Muir
>         Attachments: LUCENE-5694.patch
>
>
> These things scored implicitly on next()/advance() for historical reasons: but now that minShouldMatch is split apart we should remove this, its unnecessary.
> This allows us to remove the horrendous afterNext and clean things up, too.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org