-
andy_seaborne committed revision 7053 to the Jena SVN repository, changing 1 files.
2009-12-31 12:50:16 UTC by andy_seaborne
-
andy_seaborne committed revision 7052 to the Jena SVN repository, changing 1 files.
2009-12-30 23:02:35 UTC by andy_seaborne
-
andy_seaborne committed revision 7051 to the Jena SVN repository, changing 3 files.
2009-12-30 23:02:17 UTC by andy_seaborne
-
andy_seaborne committed revision 7050 to the Jena SVN repository, changing 2 files.
2009-12-30 22:55:26 UTC by andy_seaborne
-
andy_seaborne committed revision 7049 to the Jena SVN repository, changing 3 files.
2009-12-30 20:49:45 UTC by andy_seaborne
-
andy_seaborne committed revision 7048 to the Jena SVN repository, changing 7 files.
2009-12-29 20:11:08 UTC by andy_seaborne
-
andy_seaborne committed revision 7047 to the Jena SVN repository, changing 12 files.
2009-12-27 21:35:08 UTC by andy_seaborne
-
Summary: this affects OntModels, not plain models. The MultiUnion reifier of is the reifier of the base which is the 0'th graph of the union. This gets called twice.
Analysis:
The OntModel has a graph structure which is a MultiUnion of one graph.
Graph.find calls down to
==> GraphBase.find(TripleMatch m)
which does:
return reifierTriples( m ) .andThen( graphBaseFind( m ) ); }.
2009-12-27 20:25:08 UTC by andy_seaborne
-
With reification, can get duplicate triples from a find.
2009-12-24 14:24:55 UTC by andy_seaborne
-
This happens because javacc silently truncates files when bad UTF-8 data is found during it's buffered lookahead. Because it's early in the file, the parser actually sees no characters, hence the empty model.
There isn't much that can be done about this without reading the data twice. The behaviour is hardwired into javacc.
The Turtle parser in TDB does not have this problem and reports...
2009-12-24 14:21:31 UTC by andy_seaborne
