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.
- 12 Sep, 2018 1 commit
-
-
Martin Furmanski authored
-
- 11 Sep, 2018 2 commits
-
-
Martin Furmanski authored
-
MishaDemianenko authored
-
- 10 Sep, 2018 7 commits
-
-
Pontus Melke authored
-
Pontus Melke authored
SemanticIndexAcceptanceTest is yet again flaky - it started just a week ago and I suspect there is an actual bug hiding here somewhere. However the shrinking done by scalacheck here hides the actual error which makes it even trickier to investigate. This PR just disables the shrinking in order to make the investigation easier.
-
Pontus Melke authored
The rewriter `mergePredicates` tries to be clever by merging predicates when possible, e.g. - `a.prop IN [1,2,3] OR a.prop IN [4]` -> a.prop IN [1,2,3, 4]` - `a.prop IN [1,2,3] AND a.prop IN [2,3,5]` -> a.prop IN [2,3]` However it did the wrong thing and also rewrote expressions like - `(a.prop IN [1] AND FALSE) OR a.prop in [2]` -> `(a.prop IN [1,2] AND FALSE) AND a.prop IN [1,2]` which is not equivalent, e.g. `a.prop=1` will now evaluate to `true` instead of `false`. Fixes #12029 changelog: Fix [issue](https://github.com/neo4j/neo4j/issues/12029) where combining multiple `a IN` with AND and OR could lead to the wrong result.
-
Pontus Melke authored
-
Pontus Melke authored
-
Pontus Melke authored
Temporal already had a Field enum, remove the newly created one and join with the others
-
Pontus Melke authored
-
- 06 Sep, 2018 5 commits
-
-
Andrei Koval authored
-
Andrei Koval authored
At the moment all errors occured in worker threads are effectively ignored and do not fail the test
-
Oskar Hane authored
-
Oskar Hane authored
-
Oskar Hane authored
-
- 05 Sep, 2018 11 commits
-
-
Andrei Koval authored
-
Andrei Koval authored
-
Andrei Koval authored
-
Andrei Koval authored
-
lutovich authored
-
lutovich authored
-
Anton Persson authored
-
Anton Persson authored
This fixes a problem where Durations breaking Duration limit where created.
-
Anton Persson authored
-
Anton Persson authored
-
Anton Persson authored
When comparing DurationValues we convert months, days and seconds to avgLengthInSeconds. If values are to large, this calculation will overflow and return invalid comparison results. We now enforce the size limitation on DurationValue in constructor. The size limit is Long.MAX_VALUE seconds, which is +292 billion years.
-
- 04 Sep, 2018 14 commits
-
-
Ali Ince authored
-
Ali Ince authored
-
Ali Ince authored
-
Ali Ince authored
-
Ali Ince authored
-
Ali Ince authored
-
Ali Ince authored
-
lutovich authored
Make bookmarks returned through Bolt contain committed transaction IDs, instead of closed transaction IDs. A transaction is considered to be closed by the kernel when it is appended to the transaction log and applied to the store. Transaction ID is allocated when appending to the log. Transactions might be closed out-of-order of their IDs. "Last closed" ID is the highest gap-free transaction ID, which might be much lower than the ID of the just committed transaction. Including such ID in a bookmark might make the bookmark represent some past transaction and not the just committed one.
-
Simon Svensson authored
-
Simon Svensson authored
-
Simon Svensson authored
-
Simon Svensson authored
-
Simon Svensson authored
-
Simon Svensson authored
-