[ 
https://issues.apache.org/jira/browse/LUCENE-2771?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Uwe Schindler updated LUCENE-2771:
----------------------------------

    Attachment: LUCENE-2771.patch

Updated patch after commit of subissue LUCENE-2772.

Some throughts: The cache currently dont support reopening readers, as 
FilterIndexReader throws UOE on reopen (which is fine for most cases). But for 
this reader we should support reopen and implement it in the FilterIndexReader 
with optimized norms recreation (copy over in the map only reopened segments?).

Another thing about MultiNorms: We are inconsistent now: We are using 
MultiFields everywhere in core queries but not MultiNorms. E.g. for a TermQuery 
you can currently get a Scorer, but as soon as this scorer requests norms, it 
will throw UOE. We should be consistent. As we have now the 
SlowMultiReaderWrapper, we should remove MultiFields support from everywhere 
else in core (Filters and Queries, but also FieldCache?). *+1 for that from my 
side!*

> Remove norms() support from non-atomic IndexReaders
> ---------------------------------------------------
>
>                 Key: LUCENE-2771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-2771
>             Project: Lucene - Java
>          Issue Type: Improvement
>            Reporter: Uwe Schindler
>             Fix For: 4.0
>
>         Attachments: LUCENE-2771.patch, LUCENE-2771.patch, LUCENE-2771.patch, 
> LUCENE-2771_needsCache.patch
>
>
> Spin-off from LUCENE-2769:
> Currently all IndexReaders support norms(), but the core of Lucene never uses 
> it and its even dangerous because of memory usage. We should do the same like 
> with MultiFields and factor it out and throw UOE on non-atomic readers.
> The SlowMultiReaderWrapper can then manage the norms. Also ParallelReader 
> needs to be fixed.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to