[ 
https://issues.apache.org/jira/browse/LUCENE-3089?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13149580#comment-13149580
 ] 

Robert Muir commented on LUCENE-3089:
-------------------------------------

I just don't think it should be blanket policy without thinking things thru.

for example: lots of code you see on the internet opens a new indexreader for 
every search and closes it

should we seriously encourage this?! If someone seriously needs to do this, 
thats an expert case and
they can use try + finally and close themselves.

So for example, there I think it makes sense for IndexReader to not support 
AutoCLoseable, and separately
to remove the stupid IndexSearcher(Directory) so that IndexSearcher only takes 
IndexReader, so its *always*
a thin wrapper like we claim it is (which is an outright lie today). Then 
IndexSearcher would implement [Auto]Closeable
since its cheap.
 
                
> CachingTokenFilter can cause close() to be called twice.
> --------------------------------------------------------
>
>                 Key: LUCENE-3089
>                 URL: https://issues.apache.org/jira/browse/LUCENE-3089
>             Project: Lucene - Java
>          Issue Type: Bug
>            Reporter: Robert Muir
>
> In LUCENE-3064, we added some state and checks to MockTokenizer to validate 
> that consumers
> are properly using the tokenstream workflow (described here: 
> http://lucene.apache.org/java/3_0_3/api/core/org/apache/lucene/analysis/TokenStream.html)
> One problem I noticed in 
> TestTermVectorsWriter.testEndOffsetPositionWithCachingTokenFilter is that 
> providing a CachingTOkenFilter directly will result
> in close() being called twice on the underlying tokenstream... this seems 
> wrong.
> Some ideas to fix this could be:
> # CachingTokenFilter overrides close() and we document that you must close 
> the underlying stream yourself. I think this is what the queryparser does 
> anyway.
> # CachingTokenFilter does something tricky to ensure it only closes the 
> underlying stream once.

--
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: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org

Reply via email to