[SCM] librdf-trine-perl Debian packaging annotated tag, rdf-trine-0.110, created. rdf-trine-0.110
Gregory Todd Williams
greg at evilfunhouse.com
Fri Jun 7 14:21:06 UTC 2013
The annotated tag, rdf-trine-0.110 has been created
at dc0eab6efe89d7d853ee16b77f4ef9273fb2dd30 (tag)
tagging b0db160754595e76d4ee20781b3819c805cded57 (commit)
replaces rdf-trine-0.109
tagged by Gregory Todd Williams
on Mon Mar 16 00:34:25 2009 -0400
- Shortlog ------------------------------------------------------------
RDF::Trine 0.110.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (Darwin)
iD8DBQBJvdbahPK6VMqoyC0RAm1vAJwP6f7eOVWsTdNHjqU20NJ5Vs1nrgCeNJZE
OK9+xaNhDTMHo+2wgvMuxRE=
=hn5Q
-----END PGP SIGNATURE-----
Gregory Todd Williams (126):
- Fixed error reporting in bin/parse.pl SPARQL parser.
- Updated VERSION variables for RDF::Trine 0.109 (forgot to do this before the release).
- Fixed SSE serialization bug in ::Algebra::Sort.
- rdf_store_add_file.pl now relies on RDF::Trine's native RDF/XML parser (not the Redland 'guess' parser).
- Updated server.pl to use environment variables for configuration aligned with RDF::Query.
- Updated costmodel-logging.t expected costs.
- Fixed bug in t/23-model_bridge.t to allow two models from the same model class to be used in testing.
- Added (abstract) base class for RDF::Trine::Store.
- Added new RDF-Trine-Store-Hexastore package, implementing the triplestore index described in [Weiss08].
- Added new methods to seperate index-structure code from algorithmic code (in add_statements, remove_statements, etc.)
- Added node_values and node_count methods and updated code to not assume node lists are implemented with arrayrefs.
- Added RDF::Query::Model::RDFTrine::BasicGraphPattner::graph() method.
- Fixed node_values() to return the empty list if an undef index is provided.
- Added RDF::Trine::Store to MANIFEST.
- Added bin/graph-bgp.pl to produce a png of a query's BGP variable connectivity graph.
- Added t/plan-rdftrine.t to test QEP generation optimized for RDF::Trine BGPs.
- Updated ::Iterator::Bindings::as_string to return a scalar (had been returning a list).
- Added examples/query.pl to show a simple example of loading data and executing a query.
- Added tests to t/plan.t for QEP sse serialization.
- Added POD to RDF::Trine::Model::RDFS.
- Added POD description to RDF::Trine.
- Removed RDF::Query::Federate and RDF::Query::CostModel::Naive dependency in bin/graph-bgp.pl.
- Added ::Model::node_count method.
- Added RDF::Query::BGPOptimizer implementing a basic optimizer for basic selectivity-based join ordering.
- Added ::CostModel::Counted for backends that support efficient counting of triple patterns.
- Fixed ::Hexastore::node_count to return 0 when no terminal node list is found.
- Fixed test plan for t/optimizer.t in cases where no appropriate model is available.
- Added new SPARQLP feature for 'NOT' blocks for negation queries.
- Updated people's names in test data.
- RDF::Trine::Store::get_pattern now passes through optional @args (used for order-by).
- Starting doc changes for a 0.110_01 beta release.
- Underscored some methods to make it clear that they are private.
- Updated VERSION variables to 0.110_01.
- Added RDF::Query::Node::compare for sorting (either Trine or Query) node objects.
- Added ::Store::Hexastore::_seen_nodes for listing all nodes used in triple data.
- Added new RDF::Trine::Node::Literal::XML package for LibXML-backed literal node objects.
- Updated VERSION for RDF::Trine::Node::Literal::XML.
- Added POD to ::Node::Literal::XML methods.
- ::Node::Literal::new now calls ::Node::Literal::XML::new if available and applicable.
- Added initial code for implementing the Hexastore backend in C/XS.
- Split hexastore C code into seperate files.
- Added parse script to load RDF and print memory usage.
- Added terminal iterators.
- Updated hx_iter_current() to use terminal iterators.
- Updated hx_vector_iter_current to return current rdf_node.
- Added debugging flags to Makefile.
- Added #includes to hexastore_types.h.
- Turned off -Werror (unused variable warnings were preventing progress)
- Added hack-ish memory counting allocator for the AVL code.
- Moved Makefile to allow a Makefile.PL to autogenerate one.
- Added Makefile.PL, HexastoreXS.xs and typemap for initial perl integration.
- Implemented full hexastore indexing (instantiating all six index objects).
- Updated parse.c to use full hx_hexastore (instead of a single hx_index).
- Updates to main.c.
- Fixed bug in hx_get_statements for fff triple patterns.
- Added initial node store code to map node strings to integer IDs.
- Added file serialization and parsing for index structures.
- Renamed rdf_node type to rdf_node_id in preparation for actual rdf node content.
- Added missing dependency information for hexastore_types.h to Makefile.c.
- Added optimize program to re-sort nodes in a hexastore based on SPARQL ordering.
- Fixed index selection bug in hx_get_statements.
- Commented out debugging in hexastore.c.
- Renamed rdf_node_id to hx_node_id.
- Added hx_triples_count() function.
- Removed Makefile.PL and renamed Makefile.c to Makefile.
- Removed typemap and HexastoreXS.xs.
- Silenced unused-variable warning on index.h.
- Added hx_triple struct.
- Added -c arg to print program to display total triple count.
- Added initial code for variable bindings stucts.
- Cleaned up/silenced debugging in hexastore.c.
- Refactored hx_index_iter_next(), pulling out code to go to next vector and head items.
- Refactored hx_index_iter* functions for code that happens whenever new head or vector items are reached.
- Fixed type error and adding missing return statement in index.c.
- Node values of 0 passed to hx_get_statements now indicate ignored fields (will be distinct across the returned triples).
- Added triples_count field to head and vector structs, and code to maintain the count.
- Updated Makefile so that only parse program depends on libraptor.
- Variable bindings iterators now use hx_iter_vtable struct to delegate implementation.
- Moved hx_iter_vtable to hx_variablebindings_iter_vtable.
- Added initial code for mergejoin of two variable binding iterators (currently broken).
- Moved hx_nodemap into hx_hexastore struct.
- Forgot to add the mergejoin implementation files on last commit.
- Implemented serial mergejoin hx_mergejoin_run that now needs to be broken into a iterator-friendly coroutine.
- Fixed merge join implementation. Now materializes batches from both lhs and rhs at a time.
Merge branch 'hx-mergejoin'
- Updated public hx_* functions for new hx_node*-based APIs (moving away from public-facing hx_node_id-based APIs).
- Added hx_variablebindings_natural_join() function.
- Updated mergejoin to require a successful natural join of the variable bindings, not just matching join keys.
- Added variable bindings iterator vtable slot for functiont to test an iterator's ordering.
- Added libtap source files.
- Added TAP and test programs to Makefile.
- Added index test program.
- Added terminal and vector test programs.
- Added head test program.
- Added load.pl for pure-perl Hexastore.
- Added hexastore_query.pl for loading a hexastore data file and running a query against it.
- Added another XML test (from KjetilK).
- Added initial code for B+ trees (memory only for now, but designed for disk-backed).
- Fixed Makefile so test programs depend on their respective compiled object files.
- Refactored hx_add_triple so that threaded insertion is dealing with hx_node_ids, not hx_node pointers.
- Fixed typo in hx_new_btree_root.
- Updated parse.c to special-case the output file /dev/null for timing just the data load time.
- Silenced gcc int-to-pointer-cast, pointer-to-int-cast warnings
- Updated header files for API change requiring passing around a hx_world struct.
- Re-formalized the universal storage layer, replacing hx_world with hx_storage_manager.
- Replaced terminal ptr lists with B+-trees.
- Updated vector implementation to use B+-trees for storage.
- Updated Makefile to supply library linking args to gcc when compiling test programs.
- Added new hx_btree struct, and started moving API usage away from using hx_btree_node structs directly.
- Updated code to parameterize B+-tree branching size. Failing some tests... needs further investigation.
- Fixed B+-tree code to keep accurate parent pointers across node merge and rebalancing operations.
- Added materialization for variable binding iterators, allowing sorting.
- Updated node.c to differentiate integer and numeric literal types.
- Fixed setting of info->names and info->bindings in materialize.c.
- Added stub of new test program for variable binding iterators and joins.
- Renamed 'NOT' keyword to 'UNSAID' (SPARQLP extension).
- Added URI for 'unsaid' to the list of supported SPARQL extensions.
- Added POD to RDF::Query::Node::compare and RDF::Query::Node::from_trine.
- Fixed bug in ::Store::DBI::remove_statements causing bad SQL when all nodes are undef (remove all triples).
- Updated VERSION to 0.110_02 for second beta release.
- Updated SIGNATURE for 0.110_02 beta release.
- Added files to MANIFEST.
- Updated required version of RDF::Trine to 0.109 in Makefile.PL and META.yml.
- Updated t/algebra-func.t to run bloom filter tests conditionally based on $RDF::Query::Functions::BLOOM_FILTER_LOADED.
- Moved Digest::SHA1 from recommended list to required list in Makefile.PL.
- Updated SIGNATURE for 0.110 release.
-----------------------------------------------------------------------
--
librdf-trine-perl Debian packaging
More information about the Pkg-perl-cvs-commits
mailing list