This is fixed in r1098359. this realtime capability of instantiated is incompatible with MultiReader, because when MR initializes it computes and caches the maxDoc, which becomes out of date with respect to the InstantiatedIndexReader if you use it in this way.
The only reason the test didnt fail in trunk is because SlowMultiReaderWrapper does not do this caching, and this "top-level" norms is gone in trunk, MultiNorms just calls norms(field) and arraycopies itself. On Sun, May 1, 2011 at 11:47 AM, Apache Jenkins Server <[email protected]> wrote: > Build: https://builds.apache.org/hudson/job/Lucene-Solr-tests-only-3.x/7615/ > > 1 tests failed. > FAILED: org.apache.lucene.store.instantiated.TestRealTime.test > > Error Message: > null > > Stack Trace: > java.lang.ArrayIndexOutOfBoundsException > at java.lang.System.arraycopy(Native Method) > at > org.apache.lucene.store.instantiated.InstantiatedIndexReader.norms(InstantiatedIndexReader.java:332) > at org.apache.lucene.index.MultiReader.norms(MultiReader.java:308) > at > org.apache.lucene.search.TermQuery$TermWeight.scorer(TermQuery.java:106) > at > org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:517) > at > org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:332) > at > org.apache.lucene.store.instantiated.TestRealTime.test(TestRealTime.java:51) > at > org.apache.lucene.util.LuceneTestCase$LuceneTestCaseRunner.runChild(LuceneTestCase.java:1111) > at > org.apache.lucene.util.LuceneTestCase$LuceneTestCaseRunner.runChild(LuceneTestCase.java:1039) > > > > > Build Log (for compile errors): > [...truncated 7742 lines...] > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
