lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Robert Muir (Updated) (JIRA)" <>
Subject [jira] [Updated] (LUCENE-3580) Most 4.0 PostingsReaders don't obey DISI contract
Date Thu, 17 Nov 2011 03:34:51 GMT


Robert Muir updated LUCENE-3580:

    Attachment: LUCENE-3580.patch

patch with a test, and fixes for SimpleText and PreFlex, but i didn't fix any of the other
postings readers.

> Most 4.0 PostingsReaders don't obey DISI contract
> -------------------------------------------------
>                 Key: LUCENE-3580
>                 URL:
>             Project: Lucene - Java
>          Issue Type: Bug
>    Affects Versions: 4.0
>            Reporter: Robert Muir
>         Attachments: LUCENE-3580.patch
> While trying to do some refactoring, I noticed funky things going on with some codecs.
> One problem is that DocIdSetIterator says the following:
> {noformat}
> Returns the following:
>    * <ul>
>    * <li>-1 or {@link #NO_MORE_DOCS} if {@link #nextDoc()} or
>    * {@link #advance(int)} were not called yet.
> {noformat}
> But most 4.0 Docs/DocsAndPositionsEnums don't actually do this (e.g. return 0). instead
> are relying on Scorers to 'cover' for them, which is inconsistent. Some scorers actually
> upon this behavior, for example look at ReqExclScorer.toNonExcluded(), it calls docId()
on the
> excluded part before it calls nextDoc()
> So we need to either fix these enums, change these enums to not extend DocIdSetIterator
(and redefine
> what the actual contract should be for these enums), change DocIdSetIterator, or something
> Fixing the enums to return -1 here when they are uninitialized kinda sucks for the ones
summing up
> document deltas...

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:!default.jspa
For more information on JIRA, see:


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message