[dna-issues] [JBoss JIRA] Commented: (DNA-630) Executing a query resulted in a "read past EOF" error

[ https://jira.jboss.org/jira/browse/DNA-630?page=com.atlassian.jira.plugin... ] Randall Hauch commented on DNA-630: ----------------------------------- This was found while trying to fix DNA-613. I logged this because this was so tricky. LuceneSearchEngine works by building up a (complex) Lucene Query object and executing that, then collecting the tuples from the Lucene documents that are returned in the results. In this case, the AQM query included a constraint (among others) that LuceneSearchEngine converted into a custom HasValueQuery class (which extends Query). HasValueQuery creates a custom ExistsScorer class (which subclasses org.apache.lucene.search.Scorer), and the problem was that the 'nextDoc()' method was implemented incorrectly. It uses an internal counter to walk through the documents (handed to it in the constructor), and was not (correctly) checking whether the docId was less than the maxDoc, which is obtained from the IndexReader.maxDoc(). Essentially, the ExistsScorer was iterating through more documents than in the reader, and thus the "read past EOF". (The IndexReader uses a random-access mechanism to seek to the desired document, but if the document ID is too big, the seek goes beyond the EOF.) I checked the other Query subclasses in org.jboss.dna.search.lucene.query, and found a few other classes that were also implemented incorrectly. These were fixed as well, and all tests pass. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira