[
https://issues.apache.org/jira/browse/NUTCH-787?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12830900#action_12830900
]
Dawid Weiss commented on NUTCH-787:
-----------------------------------
The failing test in TestIndexSorter is caused by the change of implementation
inside Lucene. In Lucene 2.9, SegmentMerger calls IndexReader#document(int,
FieldSelector), but in 3.0 this has been changed to a call to document(int):
Document doc = reader.document(docCount);
Now, IndexSorter in Nutch overrides both methods and delegates to the
superclass (IndexReader) with mapping from old ids to new ids, but IndexReader
re-delegates back to the overriden method, so IDs are effectively remapped back
to original values.
> Upgrade Lucene to 3.0.0.
> ------------------------
>
> Key: NUTCH-787
> URL: https://issues.apache.org/jira/browse/NUTCH-787
> Project: Nutch
> Issue Type: Task
> Components: build
> Reporter: Dawid Weiss
> Priority: Trivial
> Attachments: NUTCH-787.patch
>
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
