[ 
https://issues.apache.org/jira/browse/LUCENE-5444?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13901188#comment-13901188
 ] 

Simon Willnauer commented on LUCENE-5444:
-----------------------------------------

Hey Britta, thanks for opening this issue! I actually think the patch looks 
pretty good though. Could you move the testcase into _MemoryIndexTest_ instead? 
We usually use a utility method to create a new directory & index writers in 
tests like this: 

{noformat}
 Directory dir = newDirectory();
 MockAnalyzer mockAnalyzer = new MockAnalyzer(random());
 IndexWriter writer = new IndexWriter(dir, newIndexWriterConfig(random(), 
TEST_VERSION_CURRENT, mockAnalyzer)); 
{noformat}

I also think from looking at it the test must fail since you are not closing 
the index writer, index reader and the directory. I guess you should move most 
of it into a single method and only keep the _compareTermVectors_ in a sep. 
method. Additionally I think since your patch allows it we should also test the 
`offsetGap` could you add another test for this as well? If you do so feel free 
to add a _getOffsetGap(String)_ method to _MockAnalyzer_ so you can reuse your 
tests!

> offsets in MemoryIndex broken when adding field with more than once
> -------------------------------------------------------------------
>
>                 Key: LUCENE-5444
>                 URL: https://issues.apache.org/jira/browse/LUCENE-5444
>             Project: Lucene - Core
>          Issue Type: Bug
>          Components: core/index
>    Affects Versions: 4.6.1
>            Reporter: britta weber
>              Labels: easyfix
>             Fix For: 5.0, 4.7
>
>         Attachments: LUCENE-5444.patch
>
>   Original Estimate: 168h
>  Remaining Estimate: 168h
>
> When fields with the same name are added more than once to MemoryIndex, the 
> offsets of the previous additions of the field do not seem to be taken into 
> account. As a result, MemoryIndex cannot be used for example with the vector 
> highlighter. 



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)

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

Reply via email to