Releases: orientechnologies/orientdb
3.2.36
Here it is another patch release, with a good set of fixes mostly around the query engine,
this also include a fix on the client API around refreshing client side document version when the document is updated
on the server side with a query, this should reduce the case of concurrent modification exception.
Changes
Core
- Make sure to correctly evaluate return statements in insert and update statements
- Correct rename of cluster to use every time lower case for low level structures
- Re-add support for '&&' operator for lucene geo overlap operations in the new query engine
- Removed multiple records load when checking if a result is a element
- Handled correctly in condition with collections of collections, issue #10303
- Re-add support of
@type,@sizerecord attributes and fixed@version, issue #10332 - Minor fix in script executor retry logic
Remote
- make sure to refresh the version of the in context record in the client when updated with a server side query
Gremlin
- Minor fix to guarantee compatibility with future jdk and groovy compatibilty
Artifacts
orientdb-community-3.2.36.tar.gz
orientdb-community-3.2.36.zip
3.2.35
This is mostly a usual patch release, with not too many changes, the only highlight of this release is that now a direct insert into a edge class with insert into E will fail, this should not affect much the applications that should already use create edge E from ... to ... to insert edges, in any case before upgrade run your validation tests.
Changes
Core
- Corrected regression of support of fetch plans in the function toJson in query engine
- Make sure to allow dot notation and square brackets for aggregation functions in query engine
- Minor fix on index creation query to allow not case sensitive collate names
- Avoid to use index with partial range for indexes not supporting null values
- Add support of comparison operators other than equal in chained indexes
- Block insert of edges with insert operations
- Auto convert properties with map value to documents when specified in the schema
- Corrected create link statement to support link to class
Artifacts
orientdb-community-3.2.35.tar.gz
orientdb-community-3.2.35.zip
3.2.34
Not too big release mostly fixes around the query engine, even though there are
stability changes also around the distributed environment contributed by @timw , and other corrections here and there.
Changes
Core
- Minor correction on value conversion for comparison in queries
- Corrected implementation of tracing executor to trace source properly
- Minor fix on storage open failure
- Correct set of target class when query target a cluster
Client
- Minor fix on retrying next host from client
Distributed
- make sure to ignore empty response from distributed requests
- Refactor distributed configuration management to be handled by distributed context
Server
- Minor fix on wait for shutdown
Artifacts
orientdb-community-3.2.34.tar.gz
orientdb-community-3.2.34.zip
Contributors
Assets 2
3.2.33
This release include mostly fixes in the query engine, for both correctness and performance improvement
trough use of indexes in additional use cases, it also deprecate some APIs that will be removed in the
next major release.
Changes
Core
- Marked methods deprecated for removal in the next major
- Make sure to handle record with missing id (lightweight edges) in delete query, issue #10239
- Re-add support of indexes for between queries in new query engine
- Return every time the correct document instance in APis and queries, issue #10280
- Re-add fetch from variable for traverse in the new query engine implementation, issue #10281
- Avoid to use index when the condition is empty list
- Make sure to respect sorting on IN conditions with indexed fields
- Handle correctly lightweight edges in save
- Support of indexes in 'is null' operator
- Corrected ORestricted case for count operations in new query engine
Graph
- Make sure to bind g variable for groovy-gremlin scripts
Server
- Update the root password if env variable is set, issue #10273
Artifacts
orientdb-community-3.2.33.tar.gz
orientdb-community-3.2.33.zip
3.2.32
This released include a small set of fixes and the upgrade to the latest gremlin release,
some fixes have been pulled from downstream forks and are made by
@andrii0lomakin and @leostryuk, also a pull request was made by @timw with the Lucene fix.
Changes
Core
- Fix bug in SQL engine variable name resolution
- Keep order for asSet() function
- Minor optimization in query engine class/cluster filtering
- Make sure to load record from database all the times by query engine
- Make sure that toJson of list of ids get expanded, issue #10264
Lucene
- Fix IndexReader ref leak when Lucene index modified and read in transaction
Security Module
- Removed copied code in favor of dependency
OrientDB Gremlin Driver
- Update gremlin to the latest release 3.7.2
Artifacts
orientdb-community-3.2.32.tar.gz
orientdb-community-3.2.32.zip
Contributors
Assets 2
3.2.31
Changes
Core
- Reduced logging for block size detecting
- Add support of empty array as from target of query
- Make sure to run let query every record when there is a context variable
- Make sure to be able to start and load databases on startup also when not run in distributed mode
Tools
- Add skipping of failed record on import, improved import logging add content of failed record
- Corrected parsing of rid to allow white space around it in import
- Corrected imported to avoid to terminate in case of failure to read a record
Distributed
- Make sure to retry to install the database later if it fail to install the first time
Artifacts
orientdb-community-3.2.31.tar.gz
orientdb-community-3.2.31.zip
3.2.30
Changes
Core
- Fix disk compatibility on some specific CPU architectures
Tool
- Add Support of preserving record ids on database import
Artifacts
orientdb-community-3.2.30.tar.gz
orientdb-community-3.2.30.zip
3.2.29
Changes
Core
- Fix issue on delegated records on raw access, issue #10171
- Minor fix in transaction not document record handling
- Make sure to check the security policy record exists before creating instance
Distributed
- Correct id swap revert in distributed transaction during retry
- Make sure to execute full sync in case of delta sync failure
Artifacts
orientdb-community-3.2.29.tar.gz
orientdb-community-3.2.29.zip
3.2.28
Changes
Core
- Check if the index exists before checking it's properties
- Make sure to initialize maps on update of value in query engine
- Add name to the OrientDB internal timer
- Removed optimization in query engine that cause failing to update linked records, issue #10077
- Corrected query with content of nested JSON documents to use embedded class
- Add count on the create link result set
- Make sure to call on drop on database drop
- Double Write Log is fsynced on writes. Looks like usage of SYNC flag is not enough
- Double Write Log is not truncated on file flush but only if it reaches file limit
- Double Write Log is truncated every time there are tail segments in queue
- We do not need to make fsync during Write Ahead Log truncation it will be forced by Double Write Log itself
- sync flag is removed for Double Write Log channel because we perform sync on data write
- Dirty counter is updated on write sync
Distributed
- Fixed database sync with enterprise agent
Etl
- Fix concurrent issues in not x86 architectures
Client
- removed duplicate cancel from client code
Artifacts
orientdb-community-3.2.28.tar.gz
orientdb-community-3.2.28.zip
3.2.27
Changes
Core
- Minor correction on value conversion for comparison in queries
- Corrected implementation of tracing executor to trace source properly
- Minor fix on storage open failure
- Correct set of target class when query target a cluster
Client
- Minor fix on retrying next host from client
Distributed
- make sure to ignore empty response from distributed requests
- Refactor distributed configuration management to be handled by distributed context
Server
- Minor fix on wait for shutdown
Artifacts
orientdb-community-3.2.27.tar.gz
orientdb-community-3.2.27.zip