[
https://issues.apache.org/jira/browse/LUCENE-1255?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12583794#action_12583794
]
Hoss Man commented on LUCENE-1255:
----------------------------------
FWIW: Apparently Luke gets confused by this as well.
if i'm understanding correctly, the only reason position "-1" seems to come
about is because of some internal counter that starts at "-1" since it assumes
the first token will have an increment of "1" and thus: the first position used
is "0"
Changing IndexWriter/DocumentsWriter/whatever to ignore the increment of the
first Token, or make it realitve to "0" might be a better solution in general.
cases to consider: what should it mean for position info if:
- first token has increment of 0
- first token has increment of 1
- first token has increment of 27
in my opinion, the answers should probably be: 0, 1, 27 ... but i think that
changes the common case of "1" (which would currently gets a position of "0"
right?) ... so maybe the right behavior is 0,0,26
> CheckIndex should allow term position = -1
> ------------------------------------------
>
> Key: LUCENE-1255
> URL: https://issues.apache.org/jira/browse/LUCENE-1255
> Project: Lucene - Java
> Issue Type: Bug
> Components: Index
> Affects Versions: 2.4
> Reporter: Michael McCandless
> Assignee: Michael McCandless
> Priority: Minor
> Fix For: 2.3.2, 2.4
>
> Attachments: LUCENE-1255.patch
>
>
> Spinoff from this discussion:
>
> http://mail-archives.apache.org/mod_mbox/lucene-java-user/200803.mbox/[EMAIL
> PROTECTED]
> Right now CheckIndex claims the index is corrupt if you index a Token with -1
> position, which happens if your first token has positionIncrementGap set to 0.
> But, as far as I can tell, Lucene doesn't "mind" when this happens.
> So I plan to fix CheckIndex to allow this case. I'll backport to 2.3.2 as
> well.
> LUCENE-1253 is one example where Lucene's core analyzers could do this.
--
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: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
