Apache Jackrabbit Oak 1.2.0 发布
5月23日 西安 OSC 源创会开始报名啦，存储、虚拟机、Docker 等干货分享Apache Jackrabbit Oak 1.2.0 发布，此版本现已提供下载：http://jackrabbit.apache.org/downloads.html。
Apache Jackrabbit Oak 1.2.0 增长性版本，基于并兼容 Jackrabbit Oak 1.x 版本，Jackrabbit Oak 1.2.x 被认为是稳定版本，可以用于生产环境。
Jackrabbit Oak 相比 Oak 1.0.0 的改进内容
-----------------------
新特性
CLONE - XPath queries currently require @ in front of
property names
Provide PojoSR based RepositoryFactory implementation
The storage backends should have an estimate on the
number of nodes below a given path
Similar (rep:similar) support for Solr
TarMK compaction
Debugging console
DataStore GC support for heterogeneous deployments using
a shared datastore
MBean to dump Lucene Index content and> Adding information to a? Standard MBean interface using annotations
TarMK Cold Standby
Introduce Password Expiry With Max Password Age and On First Login
Add path exclusion to JackrabbitEventFilter
Utility js methods to manage Oak data in Mongo
Add command to dump Lucene index in Oak Console
Use separate Lucene index for performing property> _lastRev recovery tool
> Persistent cache for the DocumentNodeStore
Support for atomic counters (non-clustered)
Enable support for NodeType based indexing rules
Expose checkpoints through JMX
Support index time Aggregation of repository nodes
Same Lucene index to support both property restrictions
and fulltext constraints
SegmentMK consistency check
Helper> Custom scorer for modifying score per documents
oak-run debug should list a breakdown of space usage per
record type
Support IS NULL based property restrictions in LucenePropertyIndex
Support IS NULL based property restrictions in the query engine
LogDumper rule to dump logs as part of system out in case
of test failure
改进
? Avoid JCR APIs calling other JCR APIs
Offline tool to repair MongoMK documents
Expose query plans
allow selecting a nodestore fixture for tests
Verify restore to revision on MongoNS
Implement optimised range queries
Improve perfomances of the OrderedIndex while insert
Configurable property exclusions for SolrIndexEditor and
SolrQueryIndex
Implement noInternal from JackrabbitEventFilter
Route find queries to Mongo secondary in MongoDocumentStore
FileDataStore inUse map causes contention in concurrent env
JCR Event Info should contain NodeType for all Events
FulltextQueryTest coverage for Solr indexer
Clarify contract of NodeBuilder.isReplaced()
Remove dependency of DocumentStore to MicroKernelException
BlobStore: improve code readability
Use in-memory DB for DOCUMENT_JDBC fixture
Use known.issues instead of commented tests
Add RestrictionProvider#getPattern(String,Set)
Enable passing of a execution context to runTest in multi
threaded runs
Avoid expanding empty property restrictions if not needed
register ContentRepository to whiteboard
Text content should not be stored as part of Index data
Enable copying of Lucene index to local file system for
read queries
Improved SolrCursor support for paged results
Consolidate Mongo client connection creation logic
OrderedIndex should serve range queries regardless of direction
add docu how to connect to Mongo w/ credentials
Avoid lock contention in Tracker.getServices()
Avoid lock contention in IndexTracker.getIndexNode()
Import of compressed wikipedia dump
Clean up dangling references to previous documents
Keep commit info for local changes in main document
Update Mongo Java Driver version to 2.12
Configurable number of rows in SolrQueryIndex
Don't create default (Scheduled)Executor when respective
instance is supplied
Set _lastRev on insert when new node is also commit root
Full text expression with * as 'field' should expand to
the 'catch all' field
DocumentNodeStore code robustness when persistence inconsistent
Solr index support for> Use SegmentMK for testing where possible
Dump help details for various command supported by oak-run
DocumentStore API: improve documentation of keys
Support collection object conversion to String[] in
ConfigurationParameters
Segment Explorer
Limit memory usage of DocumentNodeStore.readChildren()
Checkpoints> Generic operation tasks should be able to return specific results
Optimize SegmentWriter.prepare()
Configurable filter restrictions to be used in SolrQueryIndex
oak-run option to do diffs between TarMK revisions
Omit Configuration Retrieval for PreAuthContext
Use config parameters to write token property
Better cost estimates for traversal, property, and ordered indexes
Use 127.0.0.1 as default for MongoDB host
Cleanup test and execution fixtures
MicroKernelServer not usable with user specified
MicroKernel implementation
Optimize MutableTree.orderBefore for the common case
Add progress logs to MarkSweepGarbageCollector
UserAuthentication: enhance login states with> Unclear NodeStore.merge() contract
Add GlobalNameMapper(NodeState) constructor
oak-run main method should print available run modes
Introduce DocumentStoreException
Support for constraints like: foo = 'X' OR bar = 'Y'
Add Hint for selecting more performant index in
MongoDocumentStore#query
Avoid collision marker update when already present
Query: reduce debug log output
Log index names for which AsyncIndexUpdate would trigger
a full reindex
LuceneIndex should access Lucene Directory if query
involves FullTextSearch for cost calculation
Limit no of children listed with ls command in Oak Console
Improve the message in UniquenessConstraintViolation
exception to provide more details
Query engine is slow when using many "or" constraints
Avoid
in javadoc Add a way to pause the background async indexer
MongoBlobStore improvements
Add a utility method to create Solr synonym file for node
type expansions
Log read and write operations to different loggers
Add utility> Add checkpoint management to oak-run
Query: before throwing "traversed more than x nodes"
exception, log it
Compact only if needed
Make the cache> Remove ContentRepositoryImpl#getNodeStore
MemoryNodeBuilder#set only works on root builders
Tar compaction: log a warning if there is more than one checkpoint
Checkpoint list command should also provide the checkpoint lifetime
DocumentSplit should allow properties added by persistence
Query engine: disable or restrict built-in full-text engine
Provide option to use? Configuration SPI in JAAS
authentication when running within AppServer
Add timing to OrderedIndex
MBean to provide consolidated cache stats
Oak run debug option to list content from a tar file
Log details of _lastRev recovery
Add metatype info for Document and Segment services
Record time when checpoint is created in SegmentNodeStore
Improve logging in compaction and cleanup operation in SegmentMK
Improve logging in MongoDocumentStore
Oak Upgrade should depend on oak-jcr with a 'test' scope
Change NodeStore#release method to return a boolean
Add test for observation events when renaming a node
IndexStatsMBean should expose the time upto which the
repository state has been indexed
Review padding for blobs collection
DocumentNodeStore: Error calculating the machine> Turn async indexer checkpoint warning log to debug
Explorer should display the template> reference tracking report
AggregateIndex should support AdvanceQueryIndex
Simplify BranchStates
Integrate Lucene logging with Slf4j
Scalability benchmarks for search
Reduce usage of _lastRev
Lucene: improved batching and logging
Lucene: not using the path restriction can speed up queries
NodeObservor should catch exception in case event processing
UserAuthentication: Allow Password Change Via
SimpleCredentials Attribute
Fail for ACEs created for the admin principal
mk.getRevisionHistory: clarify since parameter
MicroKernelIT: clean test content
Pluggability of UserAuthenticationFactory
Allow specifying custom attributes in IndexPlan
Support for rep:spellcheck
Support for using query engine for search suggestions
Configurable Analyzer in Lucene index
Lucene index boosting on certain fields
Rep:similar support in Lucene index
Solr default rows number is too high
Create whitelist / blacklist of properties to be indexed
/ filtered for Solr index
TarMK cold standby: update OSGi config example files
Support for like query for string
Print tar file graph in segment explorer
Make blobSize in OakDirectory configurable
Allow native query to be run against specifc property index
Add configuration options for ldap connection pools
The unbound connection pool does not verify if the
connection is still alive
Create default repository in OakRepositoryFactory
LIRS cache: improved concurrency when using the cache loader
Use daemon threads for scheduled executor service in RepositoryImpl
Allow SegmentNodeBuilder to provide child builders of the same type
Log path for text extraction error on the Lucene index
Changing the query traversal limit should affect already
started queries
Provide a way for Index implementation to determine reindex mode
Execution Stats for async indexing
Add UserIdCredentials to Oak user management extension
Support property existence query (for Lucene)
Support native queries for non-full-text indexes
UserImporter should always set the rep:authorizableID
make generic VersionGCSupport perform better
Index node only if configured property changed
Add metadata about the changed value to a
PROPERTY_CHANGED event on a multivalued property
XD Logo for Oak Apache
ConflictHandler for merging jcr:lastModified
Associate user defined values with checkpoint
Use ConcurrentUpdateSolrServer for remote updates
Add checkpoint management for MongoMK
Update sql2.txt test to account for name property
presence in non test nodes
Simplify query debugging with MDC
Add jsr305 annotations to Solr index code
Not log very common queries
Path restriction evaluation support within LucenIndex
Provide API to create Tree from NodeState
Persistent cache: config option for autoCompact
Better handling for external binaries in the segment explorer
Mark checkpoints with> Provide details in log when reindexing is completed by
AsyncIndexUpdate
Remove dependency to MapDB
Remove option to disable background operations
Debug logging for merge failures
SegmentMK startup slow with large journals
Inspect any revision with SegmentMK explorer
LucenePropertyIndex should support pure nodeType based query
Use approx counters property index costs even when path
restriction is available
Clean up collision markers
Permission validation with Node.setPrimaryType
Pass over name and definitions to StoreStrategy.update
Improve performance of user query with scope set
Property name regex should only be applied on property name
MBean to expose DocumentNodeStore runtime state
Remove entryCount from NodeType Index
Refactor node store tests
TARMK Cold Standby improve test cases for FSDS mirroring
Move FixturesHelper into common
add a oak-run command to delete the index data in case of
inconsistencies
Abstract over repository creating in tests
Simplify injection of conflict handlers
Replace usages of Immutable* with factory calls
Provide API to create Tree from NodeBuilder
Change default blob> TarMk Cold Standby add metatype info for
'standby.readtimeout' property
Add RootFactory.createSystemRoot
Auto-refresh sessions on revision gc
Clarify Editor.childNodeChanged()
Improve logging in repository migration in upgrade
Improve logging in Lucene indexer
Allow querying on jcr:primaryType property if that
property is indexed
Enable the persistent cache by default
Do not create nodeChildren cache entry for leaf
Avoid clean up in XmlImportTest benchmark
Avoid lookup of conflicting node with
ImportUUIDBehavior.IMPORT_UUID_CREATE_NEW
Move MongoDiffCache to mongo package
Upgrade: reduce implementation dependency
Pre-populate version storage with intermediate nodes
reduce code duplication between
Utils.getSelectedDocuments and MissingLastRevCandidateIterator
Provide support for providing custom Tika config
Optimize read of known non-existing children
Index binary only if some Tika parser can support the
binaries mimeType
Restrict the maximum number of terms that will be indexed
for a single field
Support for maxExtractLength while parsing binaries with Tika
Include index name in the log from LuceneIndexEditor
Avoid building suggest/spellcheck dictionary at query time
Add validating commit hook for login tokens
DataStoreBlobStore: Merge test
MongoDataStoreBlobStoreTest into DataStoreBlobStoreTest
Make it possible to use the PermissionProvider from
within query indexes
Improve OAK-2465 avoiding ArithmeticException> Expose mongo and db versions for reporting purposes
checkDeepHistory/fixDeepHistory/prepareDeepHistory for oak-mongo.js
add more trace-logging to mongoMk to debug performance issues
Add user information in audit events in oak
remove CachingDocumentStore interface
Provide a default Tika Config similar to JR2 config
Update SegmentMK header format definition
Index usage for "not equals" conditions
Add missing Apache Jackrabbit qualifiers to OSGi services metatype
Support using Solr index for sorted queries
Update RecordUsageAnalyser to reflect changes in segment format
allow configuring subcaches of documentnodestore individually
Spellcheck/suggest queries should return no results if no
index can handle them
Add builder to file store
Make AbstractServiceTracker.stop()> Add option to oak-run check runmode to check consistency
of binary properties
Cleanup and document security> Open indexes in IndexTracker non blocking way
Protect the persistent cache against Thread.interrupt
Include cache invalidation stats in logs> background operations
Update mongo-java-driver to 2.13.0
RepositoryManager must not register WhiteboardExecutor with Oak
Set pauseCompaction default to false
observation processing too eager/unfair under load
Provide progress indication when reindexing is being peformed
IndexCopier Error occurred while removing deleted files from Local
Invoke indexUpdate only when new Document are added in
LuceneIndexEditor
High memory consumption of CompactionGainEstimate
more (jmx) instrumentation for observation queue
expose mongo's clusterNodes info more prominently
PerfLogger for NodeObserver.contentChanged()
Cost calculation takes time with solr pings even
when not fulfilling query
Avoid indexing user/group/uuid props by default
Timeout for MongoDB query
Move out non essential task processing from
DocumentNodeStore background thread
Log document as debug message on conflict
Use message from causing exception in
DocumentStoreException.convert()
Remove DOCUMENT_MK fixture (and> SegmentIdTable too eager to refresh
Ability to disable index hints for MongoDB
Use Consolidated diff for local changes with persistent
cache to avoid calculating diff again
Persistent cache: use different files in tests
Report a full observation queue situation to the logfile
Add description annotation to
RepositoryManagementMBean#startDataStoreGC
Enable CopyOnRead feature for Lucene indexes by default
Bug 修复
? Basic same name sibling support
Failure in ObservationRefreshTest
Occasional ConcurrentFileOperationsTest failure
Copying a large subtrees does not scale as expected in
the number of copied nodes on document node stores
Aggregate Index isn't usable in an osgi environment
ConcurrentAddReferenceTest#addReferences still fails
ConcurrentAddRemoveIT, ConcurrentAddIT test failures
ObservationTest failure on Windows
MongoDocumentStore fails to report error for keys that are too long
RDBBlobStoreTest fails occasionally
Mongo: Un-/CheckedExecutionException on replica-primary crash
Node isNew() is false in case the node is removed and
added in same commit
NodeStoreServices nondeterministic with customBlobStore
Add ConflictHandler/Validator to AbstractSecurityTest
failing test
org.apache.jackrabbit.core.query.VersionStoreQueryTest.testCheckin
document atomicity of DS.update(collection, keys, update)
OrderableNodesTest fails with Document_JDBC
"like '%'" does not use an index
Unresolved conflicts in TokenProviderImpl#createToken()
Improve warning logged on concurrent Session access
RestrictionProviderImpl returns empty pattern if all
supported restrictions are set
Diff cache entry too large
Cross foreign cluster revision comparison may be wrong
SolrIndexInitializer async parameter should be boolean
Sporadic IllegalStateException in
AbstractServiceTracker.getServices
Remote Solr server should explicitly handle failures in
connecting to ZK
DocumentNodeBuilder.setChildNode() runs OOM with large tree
Stats for the async indexing don't report 'done' status
when there are no changes
Ordered index: explain plan not implemented
DiffCache not populated after cache miss
ConcurrentConflictTest fails occasionally
Sample descendant field configuration not working for root
Create a benchmark for Full text search with Solr
Incorrect TarMK graph metadata validation
LastRevSingleNodeRecoveryTest fails with Mongo
RevisionComparator my incorrectly report a revision as FUTURE
Oak#createContentRepository never unregisters some of its services
NPE in MarkSweepGarbageCollector.saveBatchToFile during
Datastore GC with FileDataStore
NodeDocument _modified may go back in time
DocumentNodeStore "lease update thread" is not a daemon thread
Empty directories not cleaned up when gc run on FileDataStore
> IllegalStateException when using "lowerCase"/"lower" on a
array property
Logging in and out many sessions leads to high memory consumption
Failure when setting up DocumentNS fixture
Default sync handler property mapping does not allow
constant properties
Potential NPE in RDBDocumentStore
Travis builds time out
Blob garbage collector deletes referenced blobs for
Jackrabbit 2.x DataStores
SegmentMK: Inefficient flat node comparisons
TarMK: Incorrect tar entry verification in recovery mode
Lucene blob> Indexes: re-index automatically when adding an index
Hourly async reindexing on an> Unnecessary invocations of LastRevRecovery when recovery
already done.
Add restore option for oak-run
OOM in
org.apache.jackrabbit.oak.jcr.OrderedIndexConcurrentClusterIT::deleteConcurrently()
Expose security> component properties
Incomplete component declaration with default action provider
Concurrent System Login: slowdown for high concurrency levels
OrderedIndexConcurrentClusterIT takes too long
PropertyIndex only considers the cost of a single indexed property
> cheaper than an OrderedIndex (or a different AdvancedQueryIndex impl)
Stale documents in MongoDocumentStore cache
Query: Incorrect cost calculation for traversal
Ordered index fails with old index content
NodeTypeIndex is not conversative enough about its cost
NodeStoreKernel doesn't handle array properties correctly
FileNotFoundException during TarMK GC
Backup: "Attempt to read external blob" error
UnmergedBranch state growing with empty BranchCommit
leading to performance degradation
TarMK compaction delays journal updates
MicroKernel.read() returns negative value
TarMK compaction can create mixed segments
Query: UnsupportedOperationException for some
combinations of "or" and "and" conditions
MAX_QUEUED_CONTINUATIONS feature not working in
EventGenerator> TarMK compaction map check should switch comparison sides
UserImporter doesn't import protected rep:passwordLastModified
Restore: "Attempt to read external blob" error
UserContext needs to be adjusted to reflect changes made
with OAK-1922
UserContext doesn't reflect new group-member structure
Oak console on MongoMK fails with NPE
Session.logout performance poor
AsyncIndexUpdate unable to cope with missing checkpoint ref
Console in Oak Run does not close the NodeStore upon exit
Performance degradation due to SessionDelegate.WarningLock
Wrong time unit for async index lease time
Duplicate observation events when using JackrabbitEventFilter
Async property index update may fail on DocumentMK with
more than 100 index changes
IndexUpdate traverses the data nodes under index nodes
Wrong values reported for OBSERVATION_EVENT_DURATION
SecurityProvider is not able to add Custom
PrincipalConfiguration in CompositePrincipalConfiguration in OSGI Mode
ContentMirrorStoreStrategy should utilize path
restriction when available
Ordered index must ignore full-text queries
TokenLoginModule can't handle case insensitive userids
Confusing debug message about old revision access
Optimize access to OrderedPropertyIndex
Accessible tree below a non-accessible parent are HiddenTree
Observation events accessibility not checked correctly
TarMk compaction can still cause repository growth
TarMK: FileStore constructor loads all entries in the journal.log
BlobGcMaxAgeInSecs is set to high for MongoBlobStore
oak run benchmark has bad imports
oak-run build fails on windows
PermissionStore doesn't reflect dynamic nature of jcr:all
NodeState view at given version is not stable with
DocumentNodeStore
XPath queries with certain combinations of "or"
conditions don't use an index
XPath queries with "order by" are never converted to "union"
Error when accessing /oak:index/rep:policy/reindex property via JCR
Incompatibility with
PropertyDefinitionImpl#getValueConstraints potentially returning null
values.
SessionMBean fails to render due to NPE in getSessionAttributes
Oak Lucene index doesn't get notified about updates when
index is stored on the file system
Setting a MV Name property does not validate the namespaces
Setting a path property with a value that contains a
unregistered namespace does not fail
getPlan() output for NodeTypeIndex doesn't indicate the
index type used
Invalid type error while importing DocView xml data
SegmentNodeStore might not create a checkpoint
TarMK CompactionMap should include more nodes
Missing privileges after repository upgrade
oak-solr-osgi uses wrong version of spatial4j
ArrayIndexOutOfBoundsException in Segment.getRefId()
Node.setProperty(String, Value) fails for binary non ValueImpls
Excessive memory usage with Oracle Java 6
Optimize orderings by date fields
Hard-coded write concern
Reindexing can include hidden nodes in the index
Range queries and> consistent with JR2
Segment corruption
ImporterImpl does not convert JCR paths to Oak paths
Prevent against integer overflow in compaction check
Improve the resilence of the OrderedIndex for dangling links
Prevent null/empty passwords in ldap provider
BlobSize in OakDirectory differs between trunk and 1.0 branch
The query engine should not return hidden nodes
AsyncIndexUpdate can leave orphaned checkpoint in case of
abrupt shutdown
RDBBlobStore failure because of missing lastmod column on
datastore_data table
> AIOOBE in Segment#toString
Oak-run add option to remove all unreferenced checkpoints
Remove the MongoGridFSBlobStore implementation
Rename FileStoreCorruptException to
SegmentNotFoundException and update the NodeStore contract
Reindex removes all nodes under index definition node
Aggregation cursor introduces doubles in the result set
Tests do not tear down repository completely
Long running merge may block other commits
BackgroundWriteTest uses> SegmentExplorer better binary display
Segment Compactor will not compact binaries > 16k
Possibility of numeric overflow in blobCacheSize in DocumentMK
Intermittent Node not found at given revision with
DocumentNodeStore
TarMK cold standby: file handle leak
empty resultset for PropertyIndex and multi-value
properties with mixed OR+AND
Indexing on large content is slow
SegmentExplorer display in memory references to tar files
TarMK cold standby: logs are too verbose
DocumentNodeStore does not detect hierarchy conflict
Ignore maven-javadoc-plugin -Xdoclint:none option
Disable TokenAuthenticationTest#tokenCreationWithPreAuth
Better error message for getPrimaryItem
Oak-run "checkpoint" and "backup" don't use memory mapped files
Only enforce uniqueness of jcr:uuid for referenceable nodes
CacheConsistencyTest may fail with IllegalStateException
Last revision recover incomplete
Non-blocking reindexing doesn't finish properly
Change Password On First Login: Wrong Config Property Description
Add/remove version label fails with restricted access on
/jcr:system
Concurrent commit during compaction results in mixed segments
Repository upgrade does not correctly update jcr:all
aggregate privileges and bits
Node.getReferences throws IllegalArgumentException
OrderedIndex should not return the whole index on
property=$value statement
Full reindexing is triggered when the IndexEditor is missing
Ordered index does not return> un-restricted indexes
Increase the threshold for warning in PathIterator
MultipleMoveTest doesn't restore ACEs properly causing
later tests to fail
Aggregate Lucene no results for multiple jcr:contains
Increase package export on
org.apache.jackrabbit.oak.plugins.segment
External events incomplete
Lucene index not created if no node is indexed
NodeStoreKernel.getNodes throws when passing filter=""
Session.getItem violates JCR Spec
User Query: Condition 'nameMatches' ignores authorizable> UUID collision check is not does not work in transient space
IndexOutOfBoundsException in
o.a.j.o.scalability.ScalabilityNodeRelationshipSuite
Query with mixed full-text, "and", "or" conditions fails
Lucene Index property definition is ignored if its not in
includePropertyNames config
Observation events accessibility check should respect
session refresh settings
PermissionStoreImpl refresh root on flush
MemoryNodeBuilder NPE on base() following root refresh
SegmentBlob equals check should verify compaction in both
directions
TarMK Cold Standby can corrupt bulk segments
UserQueryManager: Sort Ignore Case Reversed
Compaction estimation time should not depend on number of
checkpoints
mvn site and site-deploy does not work as expected
Builder does not exist error during async index update
DocumentNS may expose branch commit on earlier revision
Async indexer checkpoints not cleaned up under heavy load
Corrupt repository after concurrent version operations
Using "order by jcr:score" slows down queries by a few
orders of magnitude
Update for absent document may throw NPE
oak.removeDescendantsAndSelf() does not remove
descendants with long paths
QueryEngine should not tokenize fulltext expression by default
Incorrect recovery of _lastRev for branch commit
> NodeStoreTest.moveToDescendant fails on MemoryNodeStore
Index cost is sometimes negative
UserValidator: Detect nested authorizables introduced by
change of primary type
DocumentNodeStore.diffManyChildren() reads too many nodes
UserProvider does not install autocreated items
Wrong optimization for joins with ISDESCENDANTNODE conditions
Compaction estimation includes all data segments
DataStoreBlobStore should log exception thrown while
reading inputstream
add logging in OrderedContentMirrorStorateStrategy for
better investigation
Field boost not working if the property for indexing is
picked using aggregate index rules
LIRS cache: deadlock if a value loader access the cache
IndexPlanner does not return plan for mixin based queries
NodeDocument.getNodeAtRevision() may read too many revisions
Wrong handling of InterruptedException in BackgroundThread
Diff reads too many nodes
InitializerTest.testInitializerMongoWithoutInitialContent() fails
TarMK Cold Standby FSDS mirroring
Support comments anywhere in a SQL-2 statement
TarMK Cold Standby expose standby read timeout value
Logic for writing properties in SegmentWriter.writeNode looks off
read is inefficient when there are many split documents
NPE in DocumentNodeStore#retrieve for non existing checkpoint
Cleanup org.apache.jackrabbit.oak.kernel package
Importing a group with random node name and members fails.
Property index: setting declaringNodeTypes to an empty
array can break queries
DataStoreWrapperBackend not updated to latest extensions
added to Backend interface
o.a.j.o.plugins.segment.HeavyWriteIT fails on travis
(exit code 137)
Sporadic test failure of HeavyWriteIT.heavyWrite on Travis
SegmentReferenceLimitTestIT does not close FileStore
ConcurrentAddNodesClusterIT does not close sessions
SegmentNotFoundException when keeping JCR Value references
AnnotatingConflictHandler: jcr:primaryType must be of type Name
Possibility of overflow in file length calculation
issues with JsopBuilder.encode and .escape
Clarify handling of hidden items in ImmutableTree
TarMk Cold Standby reduce verbosity of server errors on
client timeouts
oak-lucene fails to start in OSGi
Invalid cast being performed in BlobStoreFixture
Upgrade may fail with constraint exception
int overflow with orderby causing huge slowdown
DocumentNodeStore revision GC may lead to NPE
DefaultSolrServerProvider should use ConcurrentUpdateSS
for indexing
full-text search on first level
relative node returns no result
XPath to SQL-2 conversion fails due to escaping error
MissingIndexProviderStrategy too eager to set the reindex flag
TARMK Cold Standby> IllegalStateException for ValueMap on _revisions
Lucene AND query with a complex OR phrase returns incorrect result
UpdateOp.Key.equals() incorrect
LucenePropertyIndex throws IllegalStateException: No
query created for filter when using order by jcr:score
IndexPlanner returning plan for queries involving jcr:score
Regression with Node.getPrimaryNodeType and
getMixinNodeTypes wrt Jackrabbit 2.x
NoSuchElementException thrown by NodeDocument
offline compaction (via oak-run) broken
MongoDocumentStore fails for DELETED_ONCE indexed property
Long overflow in PermissionEntryProviderImpl
Query Filter looses property constraints for multiple and
conditions for same property
oak-mongo.js does not escape special characters for
regular expressions
TokenProviderImpl.createToken should always create valid node
Node.getMixinNodeTypes can revive deleted node mixins
Remove useless configPath in embedded SolrServer provider
DataStore GC: Fix incorrect tests
Range query with incorrectly formatted date
DataStore: Cleanup tests
ACL filtering on spellchecks
algorithm with O(n!) in mongoMk rebase - not finishing in years
SolrQueryIndex blindly executes a 2nd query to check
there are no more results
SegmentDataStoreBlobGCTest failing at some places
Error while configuring analyzer by composition
SharedDataStoreUtilsTest failing on travis
Entries in _commitRoot not purged
Suggest field doesn't get stored if it has only one value
TARMK Cold Standby allow cleanup on standby instance
OakSolrConfigurationProviderService used.properties
initialization fails
Session operations null check
TARMK Cold Standby marks the sync as stopped on first error
Persistent Cache: support append-only mode
VersionGC uses way too much memory if there is a large
pile of garbage
Wrong interpretation of SegmentNodeStoreService
configuration values
Lucene index rules should be case insensitive
DiffCache is inefficient
PropertyRestriction.isNotNullRestriction does not take
"list" into account
AsyncIndex should update the run status after the merge
has completed
ServiceEvent on each session login/logout
LuceneSupportTest fails
RDB: improve memory cache handling
Row estimator should have a threshold
MultiDocumentStoreTest.testInvalidateCache failing for Mongo
Failure in one of the batch in VersionGC might lead to
orphaned nodes
Support for additional encodings needed in ReversedLinesFileReader
OakSolrConfigurationProviderService wrongly uses
ignoredProperties for usedProperties
Lucene suggester should only be updated if the index is
used for suggestions
Findbugs plugin version should depend on JDK version
oak-pojosr tests fail
GCMonitorMBean#getRepositorySize reports incorrect values
Incorrect check for log level in PerfLogger
export org.apache.jackrabbit.oak.plugins.atomic
DocumentNodeStore.dispose() may leave repository in an
inconsistent state
ObservationTest.observationDispose() restarts repository
after test finished
IndexCopier might create empty files in case of error
occuring while copying
Test failures in TarMK standby: Address already in use
Failed expectations in TarMK standby tests
Glob restriction test failures on Jenkins
LMSEstimator update resetting weights
Online Backup of segment store fails on Windows because
the journal file is locked exclusively.
? Possible null pointer dereferences in ExternalLoginModule
Blob GC throws NPE
Avoid unneeded unboxing in PropertiesUtil
DocumentNodeStore.dispatch() may pass null to NodeStateDiff
ConcurrentAddIT occasionally fail with OakMerge0001
Misleading log message from IndexCopier
ObservationTest benchmark logs warnings about IllegalStateException
Background operation lock not> DocumentNodeStore is closed
NodeStateSolrServersObserver performs complete diff
synchronously causing slowness in dispatch
Misleading traversal warning message while performing query
Wrong indexed query estimates exceed more than double the
actual index entries
Tasks
document oak:unstructured performance advantages
RdbDocumentStore should use a connection pool
Add configuration for baselining-maven-plugin
document atomicity of DS.remove(collection, keys)
Create a benchmark for Full text search
Upgrade HttClient version in oak-solr-osgi
Drop oak-mk-perf
Move JR2 specific logic from oak-run to separate module
Update parent pom.xml to latest 2.9> MissingLastRevSeeker optimized only for MongoDS
Refactor BlobStore creation logic in a BlobStoreFixture
Analyze repository growth with Lucene index on SegmentMk
Move CacheStats humanReadableByteCount to Oak Commons IOUtils
Make SolrIndex implement AdvanceQueryIndex
Oak 1.1.x : Document new features in security area
> MissingLastRevSeeker non MongoDS may fail with OOM
> Avoid wrapping of LuceneIndexProvider with
AggregateIndexProvider in tests
Evaluate support for must(must_not(x)) query with Lucene
Add Sling OSGi Mocks to parent pom
Use PerfLogger in DocumentNodeStoreBranch
Copy Jackrabbit 2 S3> Upgrade Jackrabbit dependency to 2.10.0
Deprecate ordered index
Move ProgressNotificationEditor from upgrade module to core
Retire oak-mk-remote
详细 issues 列表：https://issues.apache.org/jira/browse/OAK??
更多内容请看发行说明。
Apache Jackrabbit 是由 Apache Foundation 提供的 JSR-170 的开放源码实现..
随着内容管理应用程序的日益普及，对用于内容仓库的普通、标准化 API 的需求已凸现出来。Content Repository for Java Technology API (JSR-170) 的目标就是提供这样一个接口。JSR-170 的一个主要优点是，它不绑定到任何特定的底层架构。例如，JSR-170 实现的后端数据存储可以是文件系统、WebDAV 仓库、支持 XML 的系统，甚至还可以是 SQL 数据库。此外，JSR-170 的导出和导入功能允许一个集成器在内容后端与 JCR 实现之间无缝地切换。
页:
[1]