FYI - I found the bug causing this problem in FieldInfos which caused
some field number to be swapped if a segment had gaps in its
numbering. I will add a test for this on LUCENE-2881 and upload a new
patch tomorrow.
I think I will let my machine run tests for a couple of days before we
commit LUC
On Mon, Feb 28, 2011 at 11:04 AM, Yonik Seeley
wrote:
> It's definitely index corruption.
>
I backed out LUCENE-2881 (see revision 1075413) until this corruption
can be resolved.
-
To unsubscribe, e-mail: dev-unsubscr...@lucene.
It's definitely index corruption.
Here's the lucene document right before it's indexed:
fields = {java.util.ArrayList@2816} size = 11
[0] =
{org.apache.lucene.document.Field@2818}"stored,indexed,tokenized,binary,omitNorms"
[1] =
{org.apache.lucene.document.Field@2819}"stored,indexed"
[2] =
{or
I hit this twice running tests yesterday FYI.
On Feb 28, 2011, at 9:00 AM, Yonik Seeley wrote:
> This is interesting... looks like when we're writing out the stored
> fields, we get a byte[4] instead of a byte[8]
> for our long field. I'll see what I can dig up.
>
> -Yonik
> http://lucidimagina
This is interesting... looks like when we're writing out the stored
fields, we get a byte[4] instead of a byte[8]
for our long field. I'll see what I can dig up.
-Yonik
http://lucidimagination.com
On Mon, Feb 28, 2011 at 5:36 AM, Apache Hudson Server
wrote:
> Build: https://hudson.apache.org/
Build: https://hudson.apache.org/hudson/job/Lucene-Solr-tests-only-trunk/5420/
5 tests failed.
REGRESSION: org.apache.solr.request.SimpleFacetsTest.testFacetSingleValued
Error Message:
Exception during query
Stack Trace:
java.lang.RuntimeException: Exception during query
at org.apache.s