This project is mirrored from https://gitee.com/mirrors/neo4jsource.git.
Pull mirroring failed .
Repository mirroring has been paused due to too many failed attempts. It can be resumed by a project maintainer.
Repository mirroring has been paused due to too many failed attempts. It can be resumed by a project maintainer.
- 09 Oct, 2017 1 commit
-
-
MishaDemianenko authored
Add "default_directory_factory" feature toggle in DirectoryFactory to be able to choose between default directory factory and NIOFSDirectory for lucene indexes directory.
-
- 25 Aug, 2017 1 commit
-
-
Craig Taverner authored
Close stream returned by procedure call when result is exhausted
-
- 24 Aug, 2017 2 commits
-
-
Henrik Nyman authored
-
Henrik Nyman authored
-
- 21 Aug, 2017 2 commits
-
-
Maria Scharin authored
Clarify shell settings are only for neo4j-shell
-
Mark Peace authored
Bump neo4j version to 2.0.0
-
- 18 Aug, 2017 1 commit
-
-
Ragnar Mellbin authored
Fix description for log rotation setting
-
- 17 Aug, 2017 3 commits
-
-
mariascharin authored
To avoid possible confusion that this is also applicable for cypher-shell
-
Ragnar Mellbin authored
-
Andrew Kerr authored
Correct documentation in neo4j.conf for causal_clustering.cluster_rou…
-
- 16 Aug, 2017 2 commits
-
-
Anton Persson authored
Procedures now close acquired KernelStatements
-
Andrew Kerr authored
-
- 15 Aug, 2017 1 commit
-
-
Anton Persson authored
-
- 14 Aug, 2017 1 commit
-
-
Mark Peace authored
-
- 13 Aug, 2017 1 commit
-
-
MishaDemianenko authored
Transaction timeout monitor
-
- 10 Aug, 2017 3 commits
-
-
Anton Persson authored
-
MishaDemianenko authored
Do not schedule monitor job if check interval is 0.
-
Mattias Finné authored
Legacy indexes can be read when in read_only mode
-
- 09 Aug, 2017 5 commits
-
-
Mattias Finné authored
-
Mattias Finné authored
-
Mattias Finné authored
TransactionBoundQueryContext no longer creates new statements
-
Mattias Finné authored
Instead always uses the statement passed in. This was a problem because isDeleted methods created a new statement, but didn't close it in the end. This was both unnecessary and it created problems due to causing statements to be open longer than they should, causing e.g. index reader caching to hold on to cache readers too long.
-
Mattias Finné authored
Previously simply getting a handle to a legacy index would have the transaction think that it needed write access, by mistake. This is now changed so that first the existence of the index is checked with only read requirement and then if it doesn't gets escalated to create-and-get it.
-
- 08 Aug, 2017 1 commit
-
-
Simon Svensson authored
Updated compiler dependencies
-
- 04 Aug, 2017 1 commit
-
-
MishaDemianenko authored
Introduce kernel transaction timeout monitor thread, that will run in the background and will mark timed out transactions as terminated in addition to existent execution guard checker.
-
- 03 Aug, 2017 7 commits
-
-
MishaDemianenko authored
-
MishaDemianenko authored
-
MishaDemianenko authored
-
MishaDemianenko authored
Close statement in node relationships iterable's
-
Sascha Peukert authored
-
Simon Svensson authored
-
Simon Svensson authored
-
- 02 Aug, 2017 4 commits
-
-
MishaDemianenko authored
According to Node interface getRelationships* methods should return relationship iterable's as result. In the same time, NodeProxy methods always return ResourceIterables for corresponding calls. As result, acquired statements will never be never released and that will cause all sort of side effects with obsolete lucene index readers that will be cached till the end of the transaction. With changes introduced in this PR, the statement will be closed but only in case if iterator of provided iterable was exhausted.
-
Alex Averbuch authored
-
Anton Persson authored
Verified in test by performing two index reads with intermediate create. Second read should see result of create but will not if KernelStatement has leaked, because Lucene reader from first match will be reused for second read.
-
Alex Averbuch authored
Removed unnecessary test
-
- 01 Aug, 2017 2 commits
-
-
Mattias Persson authored
Disable duplicates tests in LuceneSchemaIndexUniquenessVerificationIT on power8
-
Simon Svensson authored
-
- 31 Jul, 2017 2 commits
-
-
Chris Vest authored
-
MishaDemianenko authored
Previously we had a sporadic failures of *WithDuplicates tests with CorruptedIndexException inside of lucene on power8 that we did not find a ay how to address, this PR will disable those tests in case if we running on that platform
-