[
https://issues.apache.org/jira/browse/LUCENE-4899?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Simon Willnauer updated LUCENE-4899:
------------------------------------
Attachment: LUCENE-4899.patch
thanks koji for looking at it. Yet, I think we shouldn't fix this in the way I
proposed it in the previous patch. I rather think if we have a single phrase
that is greater than the fragCharSize we should just not highlight that passage
at all. This is more conservative here and I think the correct thing to do
otherwise we can easily end up with phrases way bigger than the fragment char
size. I tried to simplify this BaseFragListBuilder a bit and make this entire
behaviour pluggable so folks can decide if they want to risk the size of the
fragments to explode?
koji can you take a look at this again?
> FastVectorHighlihgter fails with SIOOB if single phrase or term is >
> fragCharSize
> ---------------------------------------------------------------------------------
>
> Key: LUCENE-4899
> URL: https://issues.apache.org/jira/browse/LUCENE-4899
> Project: Lucene - Core
> Issue Type: Bug
> Components: modules/highlighter
> Affects Versions: 4.0, 4.1, 4.2, 3.6.2, 4.2.1
> Reporter: Simon Willnauer
> Fix For: 5.0, 4.3
>
> Attachments: LUCENE-4899.patch, LUCENE-4899.patch
>
>
> This has been reported on several occasions like SOLR-4660 / SOLR-4137 or on
> the ES mailing list
> https://groups.google.com/d/msg/elasticsearch/IdyMSPK5gao/nKZq8_NYWmgJ
> The reason is that the current code expects the fragCharSize > matchLength
> which is not necessarily true if you use phrases or if you have very long
> terms like URLs or so. I have a test that reproduces the issue and a fix as
> far as I can tell (me doesn't have much experience with the highlighter.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
