[jira] [Resolved] (OAK-2438) Evaluate support for must(must_not(x)) query with Lucene

["Amit Jain (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/OAK-2438?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Amit Jain resolved OAK-2438.
----------------------------
    Resolution: Fixed

Resolving as test exists and passes

> Evaluate support for must(must_not(x)) query with Lucene
> --------------------------------------------------------
>
>                 Key: OAK-2438
>                 URL: https://issues.apache.org/jira/browse/OAK-2438
>             Project: Jackrabbit Oak
>          Issue Type: Task
>          Components: oak-lucene
>            Reporter: Chetan Mehrotra
>            Assignee: Chetan Mehrotra
>            Priority: Minor
>             Fix For: 1.1.6
>
>
> As part of OAK-2434 fix we are changing following logic
> {code:java}
> @Override
>             public boolean visit(FullTextAnd and) {
>                 BooleanQuery q = new BooleanQuery();
>                 for (FullTextExpression e : and.list) {
>                     Query x = getFullTextQuery(e, analyzer, reader);
>                     // Lucene can't deal with "must(must_not(x))"
>                     if (x instanceof BooleanQuery) {
>                         BooleanQuery bq = (BooleanQuery) x;
>                         for (BooleanClause c : bq.clauses()) {
>                             q.add(c);
>                         }
>                     } else {
>                         q.add(x, MUST);
>                     }
>                 }
>                 result.set(q);
>                 return true;
>             }
> {code}
> With a logic where the query is always added to BooleanQuery without unwrapping. This might affect logic related to comment _// Lucene can't deal with "must(must_not(x))"_. Currently there is no testcase to check that condition and we are not sure how this change would affect such a check.
> So need to add a testcase to document the behaviour for such cases



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)