[Mulgara-dev] corrupted statement store?

[Life is hard, and then you die]

  Hi,

We just hit a problem while doing a whole bunch of updates ( 'delete
select's followed by 'insert') which looks like a data corruption
issue. The command

  delete select $s $p $o from <local:///topazproject#ri> where $s $p $o and $s <mulgara:is> <info:doi/10.1371/journal.pone.0000278.t004> from <local:///topazproject#ri>;

is failing with a StatementStoreException. Here's an excerpt of the
log at level debug:

  DEBUG 43:09 WrappedTuples> Getting column 2 from wrapped class org.mulgara.resolver.spi.LocalizedTuples
  DEBUG 43:09 SubqueryAnswer> Getting object 2 from variableList [$s, $p, $o]
  DEBUG 43:09 AppendAggregateTuples> AppendAggregateTuples 1833210619: getting column 2
  DEBUG 43:09 XAStringPoolImpl> findSPObject(484278) = "2007-03-13"^^<http://www.w3.org/2001/XMLSchema#date>
  DEBUG 43:09 XAStringPoolImpl> findGNode("2007-03-13"^^<http://www.w3.org/2001/XMLSchema#date>) = 0
  DEBUG 43:09 XAStringPoolImpl> findSPObject(48) = <#ri>
  WARN 43:09 StatementStoreResolver> Failed to modifyModel: [info:doi/10.1371/journal.pone.0000278.t004 http://purl.org/dc/elements/1.1/date "2007-03-13"^^<http://www.w3.org/2001/XMLSchema#date> local:///topazproject#ri]
  org.mulgara.store.statement.StatementStoreException: Attempt to remove a triple with node number out of range: 85026 337 -1 48
          at org.mulgara.store.statement.xa.XAStatementStoreImpl$Phase.removeTriple(XAStatementStoreImpl.java:1631)
          at org.mulgara.store.statement.xa.XAStatementStoreImpl.removeTriples(XAStatementStoreImpl.java:482)
          at org.mulgara.resolver.store.StatementStoreResolver.modifyModel(StatementStoreResolver.java:369)
          at org.mulgara.resolver.InternalResolver.modifyModel(InternalResolver.java:161)

Now the next log message is interesting:

  WARN 43:09 StatementStoreResolver> Note: if any of the query nodes are negative this is a known bug [ 1089540 ] INSERT/SELECT regression in sourceforge

Well, I looked at that bug and it was closed almost 2 years ago:
http://sourceforge.net/tracker/?group_id=89874&atid=591704&aid=1089540&func=detail

Unfortunately there's no indication in there of how it was fixed. So,
my question (I guess mainly to Andrae, since he closed it) is: is that
fix in the current code base, i.e. is this a new bug or the old one?
And if the latter, what was the fix?

Btw., it looks like a backup/restore may fix things, though I haven't
truly verified that nothing's missing.


  Cheers,

  Ronald