[
https://issues.apache.org/jira/browse/LUCENE-5401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13872936#comment-13872936
]
Robert Muir commented on LUCENE-5401:
-------------------------------------
+1, nice catch!
> Field.StringTokenStream#end() does not call super.end()
> -------------------------------------------------------
>
> Key: LUCENE-5401
> URL: https://issues.apache.org/jira/browse/LUCENE-5401
> Project: Lucene - Core
> Issue Type: Bug
> Components: core/other
> Affects Versions: 4.6
> Reporter: Michael Busch
> Assignee: Michael Busch
> Priority: Minor
> Fix For: 4.6.1
>
> Attachments: lucene-5401.patch
>
>
> Field.StringTokenStream#end() currently does not call super.end(). This
> prevents resetting the PositionIncrementAttribute to 0 in end(), which can
> lead to wrong positions in the index under certain conditions.
> I added a test to TestDocument which indexes two Fields with the same name,
> String values, indexed=true, tokenized=false and
> IndexOptions.DOCS_AND_FREQS_AND_POSITIONS. Without the fix the test fails.
> The first token gets the correct position 0, but the second token gets
> position 2 instead of 1. The reason is that in DocInverterPerField line 176
> (which is just after the call to end()) we increment the position a second
> time, because end() didn't reset the increment to 0.
> All tests pass with the fix.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
