Mailing-List: contact java-user-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: java-user@lucene.apache.org
Received-SPF: softfail (nike.apache.org: transitioning domain of
 adam@labkey.com does not designate 66.111.4.25 as permitted sender)
From: "Adam Rauch" <adam@labkey.com>
To: <java-user@lucene.apache.org>
Subject: Assert / NPE using MultiFieldQueryParser
Date: Sun, 24 Mar 2013 07:57:03 -0700
Message-ID: <02cd01ce289f$d8ac6190$8a0524b0$@labkey.com>
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_000_02CE_01CE2865.2C4F1030"
Thread-Index: Ac4onMfOnW61AXe+S6SR/lg3qRY1fg==
Content-Language: en-us

------=_NextPart_000_02CE_01CE2865.2C4F1030
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit

I'm using MultiFieldQueryParser to parse search queries.  I find that
certain query strings (e.g., "/study/" without the quotes) cause
MultiFieldQueryParser.parse() to throw an AssertionError, if asserts are
enabled.  In production, parse() returns a Query, but it seems to be
corrupt. using it to search my index results in an NPE.  This seems related
to regular expressions.  That query string is probably invalid regex syntax.
but shouldn't MultiFieldQueryParser to throw a ParseException in this case?

 
Here's a simple example that reproduces the assertion:

 
        // Turn on asserts

        ClassLoader loader = ClassLoader.getSystemClassLoader();

        loader.setDefaultAssertionStatus(true);

 
        try

        {

            Analyzer analyzer = new WhitespaceAnalyzer(Version.LUCENE_41);

            QueryParser parser = new
MultiFieldQueryParser(Version.LUCENE_41, new String[]{"title", "body"},
analyzer);

            Query query = parser.parse("/study/");

        }

        catch (ParseException e)

        {

            System.out.println("Syntax error, please rephrase your query");

        }

 
This produces:

 
Exception in thread "main" java.lang.AssertionError

                at
org.apache.lucene.search.MultiTermQuery.<init>(MultiTermQuery.java:252)

                at
org.apache.lucene.search.AutomatonQuery.<init>(AutomatonQuery.java:65)

                at
org.apache.lucene.search.RegexpQuery.<init>(RegexpQuery.java:90)

                at
org.apache.lucene.search.RegexpQuery.<init>(RegexpQuery.java:79)

                at
org.apache.lucene.search.RegexpQuery.<init>(RegexpQuery.java:69)

                at
org.apache.lucene.queryparser.classic.QueryParserBase.newRegexpQuery(QueryPa
rserBase.java:790)

                at
org.apache.lucene.queryparser.classic.QueryParserBase.getRegexpQuery(QueryPa
rserBase.java:1005)

                at
org.apache.lucene.queryparser.classic.QueryParserBase.handleBareTokenQuery(Q
ueryParserBase.java:1075)

                at
org.apache.lucene.queryparser.classic.QueryParser.Term(QueryParser.java:359)

                at
org.apache.lucene.queryparser.classic.QueryParser.Clause(QueryParser.java:25
8)

                at
org.apache.lucene.queryparser.classic.QueryParser.Query(QueryParser.java:182
)

                at
org.apache.lucene.queryparser.classic.QueryParser.TopLevelQuery(QueryParser.
java:171)

                at
org.apache.lucene.queryparser.classic.QueryParserBase.parse(QueryParserBase.
java:120)

                at QueryParserException.main(QueryParserException.java:21)

 
Turn off the asserts and parse() returns "successfully". but subsequent use
of that Query instance results in NPEs such as:

 
java.lang.NullPointerException

                at java.util.TreeMap.getEntry(TreeMap.java:342)

                at java.util.TreeMap.get(TreeMap.java:273)

                at
org.apache.lucene.codecs.perfield.PerFieldPostingsFormat$FieldsReader.terms(
PerFieldPostingsFormat.java:215)

                at
org.apache.lucene.search.TermCollectingRewrite.collectTerms(TermCollectingRe
write.java:58)

                at
org.apache.lucene.search.ConstantScoreAutoRewrite.rewrite(ConstantScoreAutoR
ewrite.java:95)

                at
org.apache.lucene.search.MultiTermQuery$ConstantScoreAutoRewrite.rewrite(Mul
tiTermQuery.java:220)

                at
org.apache.lucene.search.MultiTermQuery.rewrite(MultiTermQuery.java:286)

                at
org.apache.lucene.search.BooleanQuery.rewrite(BooleanQuery.java:429)

                at
org.apache.lucene.search.FilteredQuery.rewrite(FilteredQuery.java:334)

                at
org.apache.lucene.search.IndexSearcher.rewrite(IndexSearcher.java:616)

                at
org.apache.lucene.search.IndexSearcher.createNormalizedWeight(IndexSearcher.
java:663)

                at
org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:281)

                at
org.labkey.search.model.LuceneSearchServiceImpl.search(LuceneSearchServiceIm
pl.java:1160)

 
This is appearing on production deployments with reasonable (from a user's
perspective) search queries (e.g., "http://labkey.org/study/xml" without the
quotes).  I'd like to either turn off regex parsing altogether or detect the
syntax error at parse time so I can provide my standard syntax guidance back
to the user.

 
Thanks,

Adam


------=_NextPart_000_02CE_01CE2865.2C4F1030--