[
https://issues.apache.org/jira/browse/LUCENE-2036?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12774142#action_12774142
]
Simon Willnauer commented on LUCENE-2036:
-----------------------------------------
This is a tricky one. The DistanceFilter and its derivatives fill a distance
map for each document in the index. the FieldComparator you use for sorting
depends on the filter you pass to the search method. in the local environment
they are the same instances but in the remote scenario those instances are not
the same. In turn, the distance map is not filled so you get a null value which
is then auto boxed into a double value on sorting. This scenario is just not
gonna work with remote searchable.
I guess we won't fix this until spatial is cleaned up / rewritten.
Anyway, what are you using remote searchable for, what is your usecase? Did you
think about solving you problem with some other technology?
> Underlying Spatial Logic does not work in Remote Settings
> ---------------------------------------------------------
>
> Key: LUCENE-2036
> URL: https://issues.apache.org/jira/browse/LUCENE-2036
> Project: Lucene - Java
> Issue Type: Bug
> Components: contrib/spatial
> Affects Versions: 2.9
> Reporter: Joey Surls
> Attachments: LuceneSpatialTest.java
>
> Original Estimate: 1h
> Remaining Estimate: 1h
>
> I am using Lucene in a distributed setup.
> I have a remote Searcher on another machine executing the methods in
> Searcher. For regular text searches I have no problem. For the Spatial
> contrib (according to examples provided) there is underlying logic that is
> lost when the Filter and Sort get serialized. I've attached a junit test case
> that shows when I mimic Java RMI (by serializing the objects and read it
> back) I get a null pointer exception. I believe it has to do with the Sort
> (or Filter) referencing objects in the other class Filter (or Sort)
--
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: java-dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-dev-h...@lucene.apache.org
