v4.10.0

Added: System property "metricRegistryStartConsoleReporter" to start the metric registry console reporter automatically
Added: Add fetch hint to exclude edge vertex ids or edge ids from edge references
Added: Accumulo: Logging of large rows coming from Accumulo custom iterators
Added: Accumulo: Option to compress transfers from iterator to client
Added: methods to read multiple Streaming Property Values in one request
Changed: Elasticsearch: Throw better exceptions when elements are missing in missing document helper
Changed: Elasticsearch: bulk to squash multiple updates to same element into a single update
Changed: Accumulo: use a string lookup to store label strings to reduce memory usage when reading vertices with large number of edges
Changed: InMemory: Support property graph queries with FetchHints.NONE
Fixed: Accumulo: don't queue events when no event listeners are attached
Fixed: Concurrency issue in org.vertexium.metric.StackTraceTracker

v4.9.7

Fixed: Sorting issue when ElasticSearch attempts to implicitly cast Date Times to Doubles in search query string
Fixed: Elasticsearch: Streaming Property Values being indexed as "" after an update of the property

v4.9.6

Fixed: Threading issue when multiple threads attempt to deserialize a LazyMutableProperty value at the same time.
Fixed: When updating a visibility on a field multiple times, it was possible to set the value of the field to null. This issue has been fixed.

v4.9.5

Fixed: ConcurrentModificationException possible when mutliple threads are iterating over the metadata of the same property instance.

v4.9.4

Fixed: Improved detection and logging of Elasticsearch scrolls that are not closed properly.

v4.9.3

Fixed: Term aggregations that included the HasNot count and had no documents with the actual field were throwing an exception. This version now returns the proper count of documents (all of them) in the hasNot count.
Changed: Updated the validation logic in the GeoShape classes to more strictly adhere to the GeoJSON specification.
Changed: A new GeoUtils class provides helper methods for validating and automatically cleaning up GeoShape instances as well as methods that performing intersection and within checks.
Changed: Inverted the meaning of the GeoShape.within method. Previously, the code shape1.within(shape2) was asking if shape 2 was within shape 1 which is not intuitive. Moving forward, that same code means "is shape 1 within shape 2", which reads more naturally.

v4.9.2

Changed: Elasticsearch: search.indexRefreshInterval configuration to be a string to be in line with what Elasticsearch expects
Fixed: Elasticsearch 7 query string transformation was improperly stripping escape characters from fields that could not be found in the search index.
Fixed: When re-indexing an element with hidden properties in Elasticsearch, the hidden property field was not being included in the list of fields to add.

v4.9.1

Added: MetadataPlugin which allows filtering common values from being written to the data store
Added: Elasticsearch detect and close open scrolls
Fixed: Element/ElementId hashCode and equals
Fixed: Elasticsearch Bulk service race condition
Fixed: Blind writing of edges with flipped in/out vertex ids returning incorrect results

v4.9.0

Added: FetchHints.union
Added: Support for Elasticsearch 7. Switching an existing project from Elasticsearch 5 to Elasticsearch 7 will require the Vertexium data to be re-indexed. If using SSL, there is a known issue with the bulk index service in this version.
Deprecated: Support for Elasticsearch 5.
Fixed: InMemory search failing to query correctly when just the table name is provided
Fixed: InMemory concurrency issues with extended data simultaneous deletes and adds
Fixed: getExtendedData single row with multiple matching row id prefixes
Fixed: Elasticsearch limit while using scroll API
Removed: Accumulo map reduce code

v4.8.1

Added: Term aggregation result includes sum_other_doc_count and doc_count_error_upper_bound
Added: Term aggregation parameter named includeHasNotCount to include a count of the number of element missing the aggregated property. This is false by default.

v4.8.0

Added: Option to refresh on flush as opposed to flushing at query time
Added: Graph.getElements - get multiple elements (vertices and edges) with one call
Fixed: Soft delete with event data cannot be un-soft deleted
Changed: Elasticsearch to use bulk updates
Changed: EdgeVertexPairs can now have null vertex ids if user can see the edge but not the vertex
Deprecated: the following methods have been deprecated in favor of methods on mutations.
- Graph.addVertex
- Graph.addEdge
- Element.deleteProperty
- Element.setProperty
- Element.markPropertyHidden
- Element.markPropertyVisible

v4.7.1

Added: Compare.RANGE to query
Added: Support to build on JDK 11
Changed: Range renamed to IdRange

v4.7.0

Added: Element/Extended Data Row additional visibilities
Added: Sorting strategy
Added: Element#getExtendedData - without table name, to get all extended data rows for that element
Added: Fetch hints to get old metadata
Fixed: "Scanner already closed" exception in ScannerStreamUtils
Fixed: Query on Extended Data reserved field names (ELEMENT_TYPE, ELEMENT_ID, TABLE_NAME, ROW_ID)
Fixed: Accumulo decoding input streams that end with zero length byte arrays
Changed: Speed up element delete

v4.6.3

Fixed: Elasticsearch query string not taking account element visibility in total hit count

v4.6.2

Fixed: Elasticsearch query strings with field names - handling of values other than constants

v4.6.1

Fixed: Elasticsearch Compare.NOT_EQUAL and Contains.NOT_IN with multiple visibilities

v4.6.0

Added: History for vertices, edges, and properties
Changed: Accumulo use a thread pool when doing batch scans
Changed: Throw exception on bad keys/ids
Fixed: Elasticsearch deletion of geo-location properties
Deprecated: Graph.getVertexPropertyCountByValue

v4.5.2

Fixed: Accumulo delete extended data not removing rows
Changed: Elasticsearch to not use locks when flushing it's internal queue
Changed: Elasticsearch performance improvements to Hamming Distance painless script

v4.5.1

Changed: Improve Accumulo iterator memory by not converting values to strings and then back again
Changed: Cypher: update to latest openCypher tck
Removed: Ability to run embedded Elasticsearch

v4.5.0

Added: Accuracy to GeoPoint object
Fixed: If you had access an element but none of the extended data columns you could still match on the extended data row using a query with no filters. This would result in the row being counted in totalHits, but it could not be retrieved from Accumulo. NOTE: This change requires that all extended data values are re-indexed in ElasticSearch.
Fixed: If the visibility of an element changes, Vertexium will now re-index all associated extended data to ensure that they can be found according to the new visibility.

v4.4.2

Fixed: EdgeInfo.getDirection was reversed when loaded from InMemoryVertex.
Changed: Elasticsearch: index on reads instead of indexing on writes.
Changed: Elasticsearch: optimize adding properties to vertices/edges with lots of properties

v4.4.1

Fixed: Soft Deleted Historical Property Entries not containing metadata information.
Fixed: Scoring strategies selecting the correct fields
Fixed: Indexing of GeoShape properties when the property is delete or the visibility is updated.
Fixed: Soft Deleted Historical Property Entries not containing deleted value.

v4.4.0

Added: The ability to perform a term query by prefix using a new Compare.STARTS_WITH predicate.
Changed: Accumulo to lazily load metadata values
Changed: Metadata was changed to an interface. new Metadata needs to be changed to Metadata.create
Fixed: AccumuloGraphLogger was throwing an error when trying to log a field from a Scanner that was removed in Accumulo 1.8.0

v4.3.1

Fixed: Serialization of StreamingPropertyValueRef

v4.3.0

Added: Cardinality aggregation
Changed: Remove tools: Reindex, Backup, Restore
Changed: Remove FetchHintsBuilder.parse
Fixed: GeoPoint.calculateCenter
Removed: PropertyValue.store (this include streaming property values), property values are not always stored

v4.2.0

Added: Elasticsearch handler for missing documents
Added: Query edges with a given in or out vertex
Changed: Elasticsearch version to 5.6.10
Changed: Accumulo version to 1.9.2

v4.1.3

Added: Query not equals element id
Added: Query limit by minimum score
Added: Field value scoring strategy

v4.1.2

Added: Per query scoring strategy
Added: Hamming distance query scoring strategy
Added: Aggregating extended data by element id and element type
Added: Elasticsearch TransportClient plugins support (config parameter search.esTransportClientPlugin.xxx=<plugin class name>)
Added: Accumulo additional client configuration options (config parameter clientConfiguration.xxx=<value>)
Changed: Deprecated Elasticsearch global scoring strategy

v4.1.1.7

Added: Ability to pass in a timestamp when calling getVertices

v4.1.1.6

Fixed: Accumulo: Find path traversing edges that were soft deleted then re-added

v4.1.1.5

Changed: Elasticsearch: Parse histogram aggregation interval to Double since decimal intervals are supported

v4.1.1.4

Fixed: Elasticsearch: NPE in InfiniteScrollIterable if no matching property name has been found when applying filters

v4.1.1.3

Invalidate the define property cache

v4.1.1.2

Added: Query edges with a given in or out vertex
Added: Elasticsearch: Return ElasticsearchEdge and ElasticsearchVertex when FetchHints.EDGE_REF and FetchHints.NONE is set for performance

v4.1.1.1

Fixed: graph.findRelatedEdgeSummary filtering edges that were re-added after soft delete/hidden

v4.1.1

Added: Search configuration parameter search.splitEdgesAndVertices. The default value is false. If set to true, edges and vertices will be stored in separate search indices.
Added: Option to Elasticsearch index to log large document (search.logRequestSizeLimit)
Added: Ability to parse from json to FetchHintsBuilder
Changed: Elasticsearch: Timeout to flush documents from 30 seconds to 30 minutes
Changed: Disable leading wildcard support
Changed: Update to curator 2.12.0
Changed: Update to Hadoop 2.9.0
Changed: Update to Zookeeper 3.4.12

v4.1.0

Added: Edge query support for in and out vertex ids
Changed: If fetch hints were not provided throw an exception instead of returning null
Changed: Elasticsearch to use BulkProcessor
Changed: Calling hasId multiple times on the same Query object will result in the intersection of the provided lists. Previously, the resulting list was a union of all provided inputs. This behavior is more consistent with the AND behavior of the other query methods.
Fixed: Updating metadata when metadata was not fetched in fetch hints

v4.0.0

Changed: FetchHints to support more filtering
Changed: Throw errors when calling methods without the proper fetch hints
Changed: Accumulo: Store metadata in an indexed list to prevent duplication in memory and over the wire
Fixed: Accumulo Iterator memory leak
Added: Method to Vertex to get more detailed edge summary
Removed: SQL
Removed: Blueprints and Titan support

v3.2.3

Fixed: Not adding all of the geoshape fields to index for existing elements
Fixed: Expand SPVs when running in memory query strings
Fixed: Fix query search using different authorizations query string
Fixed: Fix indexing of geolocation properties
Fixed: notification to Elasticsearch when properties are added

v3.2.2

Fixed: Improved support for multithreaded clients with InMemoryGraph
Fixed: Saving BigInteger and BigDecimal when using prepareMutation with existing element

v3.2.1

Changed: When saving an ExistingElementMutation, the Elastisearch5Index will now apply the mutations using a painless script rather than making multiple requests.
Changed: When using Accumulo, all threads now share a single batch writer rather than creating a new writer for every thread. This allows client programs to more effectively use multi-threading.
Fixed: Prevent concurrent modification exceptions when using InMemoryGraph through synchronizing the methods and also returning Collection copies.
Fixed: The InputStream for a StreamingPropertyValue stored in Accumulo now supports mark/reset.
Fixed: Elasticsearch sort throwing errors for String properties while querying across indices

v3.2.0

Changed: Elasticsearch _id and _type fields to be much smaller to minimize the size of the _uid field data cache
Changed: Elasticsearch to only refresh the indices that were changed
Fixed: InMemory and Accumulo fix wrong vertex being returned after re-adding the same vertex with a different visibility after soft delete
Fixed: Find path when edge labels are deflated
Fixed: Accumulo stream property value in table data length of reference
Fixed: Query will now return hidden elements/vertices/edges if FetchHint.INCLUDE_HIDDEN is passed
Fixed: Query.hasAuthorization will now match elements whose only use of an authorization is that the element is hidden
Fixed: Query.hasAuthorization will now match elements whose only use of an authorization is a hidden property
Fixed: Elasticsearch server plugin missing fields while querying across indices
Fixed: Elasticsearch shard exception if geo property is not defined
Added: Query methods elementIds/vertexIds/edgeIds are overloaded to accept IdFetchHint, which makes it possible to include ids for hidden elements
Added: Graph.getExtendedDataInRange in order to bulk load ranges of extended data rows
Added: SearchIndex.addExtendedData to allow for reindexing extended data
Added: Elasticsearch exception if any shard failures occur on queries
Added: Elasticsearch option to force disable the use of the server side plugin

v3.1.1

Added: Graph.getExtendedData to get a single extended data row
Fixed: InMemory update extended data with different value

v3.1.0

Changed: Set Elasticsearch scroll api query to have a size that is the same as page size
Changed: Properly throw NoSuchElementException from Iterables when next is called with no more elements
Changed: If exact match is chosen for a property, full text is automatically enabled to support aggregation results
Added: Accumulo in table storage of streaming property value data
Added: Functions to delete extended data columns
Added: Multi-value extended data columns
Added: Sorting and aggregating of row ids and table names
Added: Option to execute a script from the shell
Fixed: Streaming property value input stream from data table
Fixed: Elasticsearch5 retry logic to sleep per request failure instead of batch of failures
Fixed: Mixed case searching on exact match properties
Fixed: Intermittent bug while indexing streaming property values under high load
Removed: ElasticSearch 1.x support

v3.0.4

Fixed: Infinite loop in PagingIterable if Elasticsearch returned vertices that wasn't saved in the database yet
Fixed: Pagination returning duplicate ids on different pages by adding default sort by score and then id. If sorts are specified, sort by score and then id after specified sorts. If sorts are not specified, sort by score and then id.
Fixed: Elasticsearch flush overflow causing excessive exceptions
Fixed: Fix Elasticsearch5 missing field exception when sorting on multiple indices

v3.0.3

Changed: Accumulo default data storage to use in table storage and not overflow to HDFS
Added: Added a hasAuthorization method to the Query class to allow searches for any element that uses an authorization string or strings.
Added: Query extended data on an element
Added: EmptyResultsGraphQuery and EmptyResultsQueryResultsIterable
Fixed: GeoPoint.distanceBetween and GeoPoint.calculateCenter calculations
Fixed: Elasticsearch5 field limit of 1000
Fixed: Query.hasId to work with extended data rows
Removed: CompositeGraphQuery

v3.0.2

Fixed: Elasticsearch5 throwing unsupported operation exception when adding/updating a property
Added: Added a has method to the Query class to allow searches on all properties of a particular data type.
Added: Added a has method to the Query class to allow searches for any elements with/without a property of a particular data type.
Added: Added a has method to the Query class to allow searches for a value across multiple properties.
Added: Added a has method to the Query class to allow searches for a elements with a list of properties. The presence of any property in the list will cause the document to match.
Added: Added a has method to the Query class to allow searches for a elements without a list of properties. The absence of all properties in the list will cause the document to match.
Added: Implemented support all compare operators for DateOnly field types when using Elasticsearch 5.
Fixed: Queries with both a query string and aggregations were throwing a "not implemented" exception in the Elasticsearch 5 plugin.

v3.0.1

Changed: Find Path to not return a path that contains one or more vertices that is can't be retrieved because of visibility restrictions
Changed: Reduced DefaultIndexSelectionStrategy cache load time from 1hr to 5min
Added: Added a hasId method to the Query class to allow searches to be filtered by element ID.
Fix: Extended data element type value for edges
Changed: Field removal from Elasticsearch documents is now queued as a future instead of immediate
Fix: Marking vertices/edges as hidden will now update the document in the search index as well as the data store
Fix: Fixed an infinite looping problem in the PagingIterable that may have resulted from a malformed ES document
Fix: Elasticsearch5SearchIndex.removeFieldsFromDocument will not run the removal script in Elasticsearch if there are no properties to remove
Fix: Precision and error are now configurable when mapping GeoShape properties. This allows an adjustable trade off between index performance and false positive search results.

v3.0.0

Changed: Removed ES 2 support and replaced it with ES 5 support
Changed: Removed support for NameSubstitutionStrategy from the Elasticsearch modules. NameSubstitutionStrategy is still available for Accumulo modules
Changed: Added property checking when using the has method to query the graph for a value. An exception will now be thrown if an attempt is made to either query a property that does not exist or to text query a property that isn't full text indexed
Note: Upgrading to this version will require re-indexing if you use NameSubstitutionStrategy or are switching to the ES 5 module.
Added: Support for extended GeoShape storage and search. Currently the new shapes are only supported by the ElasticSearch 5 module.
Added: GraphVisitor to visit elements, properties, extended data rows using a visitor pattern

v2.6.2

Changed: remove deprecated interfaces, examples, and changed Elasticsearch deprecated methods
Fixed: Contains.NOT_IN not returning vertices the property does not exist in
Added: Does Not Contain Text Query
Added: Elasticsearch shard size configuration for term aggregations
Added: Elasticsearch scroll API support

v2.6.1

Fixed: SQL SPV loading when timestamp is out of sync with property
Fixed: Elasticsearch plugin handling boolean properties in query

v2.6.0

Added: ability to query for vertex/edge/element ids rather than the elements themselves
Added: ability to return if there is any paths between two vertices
Added: ability to store extended data rows on an element
Added: Elasticsearch 2.x support
Added: Cypher query support
Added: CLI: profiles to allow running the CLI from within an IDE
Added: hasDirection and hasOtherVertexId to VertexQuery
Changed: The default behavior of calling methods without fetch hints will use the default fetch hints specified in the graph configuration
Changed: removed dist module
Fixed: Issue #135. Passing FetchHint.NONE when retrieving vertices from Accumulo using Elasticsearch will now properly return the vertices rather than an empty Iterable
Deprecated: EdgeCountScoringStrategy

v2.5.6

Fixed: DeleteHistoricalLegacyStreamingPropertyValueData with property keys having common prefix

v2.5.5

Changed: Removed timestamp from streaming property value row key.

v2.5.4

Fixed: Find Path max 2 hops not returning one hop paths
Added: configuration for number of Elasticsearch replicas
Changed: Delete events added to HistoricalPropertyValue reporting

v2.5.3

Added: size to TermsAggregation
Added: support for Elasticsearch percentile aggregations
Added: Elasticsearch: Field query string support

v2.5.2

Added: support for Elasticsearch range aggregations
Added: hasChanges method to mutations
Added: Edge.getVertices which accepts fetch hints
Added: support to read multiple StreamingPropertyValues with on query
Added: Graph.findPath exclusion labels
Added: methods to support deleting multiple properties on a single ES request to minimize collisions.
Added: support storing history in a separate table
Changed: avoid saving empty mutations in Graph#saveElementMutations
Changed: if the Elasticsearch write queue already contains element flush
Fixed: existing element mutations updating timestamps
Fixed: bulk request conflict by adding a retry count

v2.5.1

Changed: Speed up property reads by using Maps to directly get the property
Fixed: updated to repopulate the Metadata entriesLock in the case that the class gets deserialized

v2.5.0

Added: Optional benchmarks unit test
Changed: ElasticSearch: asynchronously submit element updates
Changed: AccumuloGraph to use MultiTableBatchWriter
Changed: org.vertexium.Metadata accesses are now protected with ReadWriteLock
Fixed: More checks for null vertex ids or labels when creating edges
Changed: Upgrade elasticsearch version to 1.7.5
Changed: Update accumulo version to 1.7.2
Added: getVertices helper method on Edge
Added: graph.saveElementMutations

v2.4.6

Changed: Delete events added to HistoricalPropertyValue reporting
Add test and javadoc for altering property that has a property key that is not the default

v2.4.5

Added: Vertex.query limited by connected edge labels
Fixed: Memory leak when using compression in serializer
Speed up Accumulo EdgeInfo and fix null conditions
Add validation for edge in/out vertices and label

v2.4.4

Quick Kryo Serializer: add support to compress the bytes after serialization
InMemory/SQL: Fix historical metadata values getting lost on change of metadata
Graph: Remove the comparison of graph in the GraphEvent base class
Added interface GraphWithSearchIndex, which GraphBaseWithSearchIndex now implements

v2.4.3

Accumulo: fix accumulo not cleaning up property with old visibility in search index

v2.4.2

Accumulo/Blueprints: Exclude ripple-flow-rdf from dependencies because of transitive dependency issues
maven: Update maven plugin versions
Util: add helper method that makes it a little cleaner to check for an empty Iterable without having to worry about closing the iterator.
InMemory/SQL: update the InMemoryVertex implementations of getVertices and getVertexIds in order to make it behave consistently with the Accumulo implementation. When faced with edges to Vertices that can't be seen, getVertices will omit them and getVertexIds will return the ids

v2.4.1

InMemory/SQL: improved find paths performance
InMemory/SQL: fix in-memory vertex properties inadvertently sharing metadata
Graph: remove excessive warning when scanning all elements.
Graph: add methods to get vertices and edges in a range of IDs
Accumulo: add methods to get the tablet splits
Elasticsearch: Add extended bounds to histograph aggregation

v2.4.0

ACCUMULO: iterator locations in accumulo config are now stored per table name
Elasticsearch: fix: sort by strings with tokens should not effect sort order
SQL: Switch to HikariCP connection pool

v2.3.1

Elasticsearch: fix: calendar date field aggregations with multiple visibilities
Elasticsearch: Support additional configuration for in process node
Elasticsearch: fix: Aggregation after alter visibility
SQL: fix: refresh in memory representation after altering vertex visibility

v2.3.0

SQL: index vertex columns on edge table
Elasticsearch: Load field mapping on startup
Search: Add CalendarFieldAggregation
StreamingPropertyValue: Add readToString method with offset and limit

v2.2.12

Elasticsearch: only sort on properties know to be in the index
fix inmemory implementation to handle soft deleting with IndexHint of DO NOT INDEX
close SPV input stream in readToString

v2.2.11

SQL: Change max primary key in DDL to 767 to fix MySQL
Elasticsearch: use an inline Groovy script to efficiently delete a property; also fixes query bugs
ACCUMULO: fix markPropertyVisible for edge properties
ACCUMULO: fix saving mutation with the save key/name/visibility properties with different timestamps

v2.2.10

ACCUMULO: fix find path traversing over deleted edges
InMemory: add hashCode() to InMemoryElement

v2.2.9

SQL: change varchar size in create tables from 100 to 4000

v2.2.8

SQL: create tables if they do not exist
add method to get all historical values of all properties of an element

v2.2.7

Elasticsearch: disable in-process zen discovery by default

v2.2.6

allow overriding ES configuration
for AccumuloGraph, when altering an element visibility also add the element to the search index with the new visibility
fix getLength in sql streaming value property

v2.2.5

add possible configuration to set hadoop conf dir

v2.2.4

term aggregation is supported in the in memory graph query
add serialVersionUID to serializable SQL classes
support no default constructors in Kryo
fix in memory getEdgeIds and getEdgeInfos when getting hidden edges to be consistant with AccumuloGraph

v2.2.3

if no matching properties are found for a property return an empty result set. See VertexiumNoMatchingPropertiesException
for SqlGraph, optimized SQL query used to retrieve vertices
fix softDeleteProperty on hidden elements
change the default search.scoringStrategy to org.vertexium.elasticsearch.score.NopScoringStrategy

v2.2.2

fix property definition bug in map reduce code
add strictTyping configuration, defaults to false for backwards compatibility
Use auto deleting file stream if the length of streaming property value is unknown

v2.2.1

support Geohash searches
helper methods on Vertex to get Vertex/Edge pairs
better ElasticsearchSingleDocumentSearchIndex retry logging
merged ElasticsearchSingleDocument and ElasticsearchBase
support edge labels in aggregations

v2.2.0

add IP Address type
support multi-level aggregations
query bug fixes

v2.1.0

Changed ValueSerializer to VertexiumSerializer. This requires a change to your configuration: accumulo.graph.valueSerializer is now accumulo.graph.serializer and the class name specified should change as well.
get multiple metadata entry values helper methods Metadata.getEntries and Metadata.getValues
CLI improvements
- delete and query shortcuts
- additional help
- Upgrade groovy to 2.4.5
- Upgrade jline to 2.13
Query
- Search for multiple edge labels
Elasticsearch
- optionally run Elasticsearch in process (i.e. no ES server needed)
- indexes label, in vertex id and out vertex id
- vertex query not filtering edges (issue #41)
- fix hasNot query when no properties are set with that value
- change the ES query to return based on element visibility not individual property visibilities
alterPropertyVisibility bug fixes
AccumuloGraph.findRelatedEdgeSummary bug fix to filter out soft deleted and hidden edges
InMemoryGraph bug fixes
sort no longer throws exceptions on field not found (issue #47)
Introduced an experimental SqlGraph for using a relational database as an alternative to Accumulo.

v2.0.3

statistics aggregation support
aggregation bug fixes

Files

CHANGELOG.md

Latest commit

History

CHANGELOG.md

File metadata and controls

v4.10.0

v4.9.7

v4.9.6

v4.9.5

v4.9.4

v4.9.3

v4.9.2

v4.9.1

v4.9.0

v4.8.1

v4.8.0

v4.7.1

v4.7.0

v4.6.3

v4.6.2

v4.6.1

v4.6.0

v4.5.2

v4.5.1

v4.5.0

v4.4.2

v4.4.1

v4.4.0

v4.3.1

v4.3.0

v4.2.0

v4.1.3

v4.1.2

v4.1.1.7

v4.1.1.6

v4.1.1.5

v4.1.1.4

v4.1.1.3

v4.1.1.2

v4.1.1.1

v4.1.1

v4.1.0

v4.0.0

v3.2.3

v3.2.2

v3.2.1

v3.2.0

v3.1.1

v3.1.0

v3.0.4

v3.0.3

v3.0.2

v3.0.1

v3.0.0

v2.6.2

v2.6.1

v2.6.0

v2.5.6

v2.5.5

v2.5.4

v2.5.3

v2.5.2

v2.5.1

v2.5.0

v2.4.6

v2.4.5

v2.4.4

v2.4.3

v2.4.2

v2.4.1

v2.4.0

v2.3.1

v2.3.0

v2.2.12

v2.2.11

v2.2.10

v2.2.9

v2.2.8

v2.2.7