On Fri, Jan 25, 2013 at 2:30 AM, Robert Muir <rcm...@gmail.com> wrote: > I think the bug is when its direct and the last block has the > optimized bitsPerValue=0 case.
Right. This was a test bug due to the fact that the reader moves the file pointer in the "for (i = 0; i < valueCount; ++i) { assert blah }" loop. I committed a fix. Thanks for running tests on this branch! -- Adrien --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org