Leverage `skip` and `limit` methods exposed in `GridFSFindIterable` to support limiting and skipping results.
In particular these changes allow the `find(Query)` method to correctly consider parameters of a page request.
Original pull request: #806.
MongoServerCondition replaces the JUnit 4 TestRules (MongoVersionRule & ReplicaSet) with a JUnit Jupiter ExecutionCondition.
Original Pull Request: #807
We now use a non-blocking state switch to determine whether to invoke drainLoop(…) from Subscriber completion.
Previously, we relied on same thread identification assuming if the subscription thread and the completion thread were the same, that we're already running inside the drain loop.
It turns out that a I/O thread could also run in event-loop mode where subscription and completion happens on the same thread but in between there's some processing and so the the call to completion is a delayed signal and not being called on the same stack as drainLoop(…).
The same-thread assumption was in place to avoid StackOverflow caused by infinite recursions.
We now use a state lock to enter the drain loop. Any concurrent attempts to re-enter the drain loop in Subscriber completion is now prevented to make sure that we continue draining while not causing stack recursions.
Original Pull Request: #807
Projections used within an aggregation pipeline can result in empty documents emitted by the driver. We now guarded those cases and skip those documents within a Flux or simply return an empty Mono depending on the methods signature.
Original pull request: #804.
AsyncInputStreamAdapter now properly splits and buffers incoming DataBuffers according the read requests of AsyncInputStream.read(…) calls.
Previously, the adapter used the input buffer size to be used as the output buffer size. A larger DataBuffer than the transfer buffer handed in through read(…) caused a BufferOverflow.
Original Pull Request: #799
We now rely on an outer drain-loop when GridFS reads complete on the same thread instead of using recursive subscriptions to avoid StackOverflow. Previously, we recursively invoked subscriptions that lead to an increased stack size.
Original Pull Request: #799
maxTimeMs defines the time limit for the aggregation operations. If not specified or set to zero, operations will not time out.
Original pull request: #800.
Execution time and test order changed by using JUnit5. This commit fixes some of the issues related to index creation where actually not needed.
Original pull request: #798.
This commit switches from simple collection.count(), operating on potentially false collection statistic, to countDocuments() using an aggregation for accurate results.
The transition required query modifications at some points because $match does not support $near and $nearSphere but require $geoWithin along with $center or $centerSphere which does not support $minDistance (see https://jira.mongodb.org/browse/SERVER-37043).
$geoWithin further more does not sort results by distance, but this fact can be ignored when just counting matches.
Examples:
{ location : { $near : [-73.99171, 40.738868], $maxDistance : 1.1 } }
{ location : { $geoWithin : { $center: [ [-73.99171, 40.738868], 1.1] } } }
{ location : { $near : [-73.99171, 40.738868], $minDistance : 0.1, $maxDistance : 1.1 } }
{$and :[ { $nor :[ { location :{ $geoWithin :{ $center :[ [-73.99171, 40.738868 ], 0.01] } } } ]}, { location :{ $geoWithin :{ $center :[ [-73.99171, 40.738868 ], 1.1] } } } ] }
Original pull request: #604.
Add author tags. Move integration tests to existing test class.
Apply more appropriate in existing tests assertions. Use diamond syntax.
Original pull request: #796.
We now consistently use GridFSFile.getId() to allow custom Id usage instead of enforcing the Id to be an ObjectId. Using the native Id allows interaction with files that use a custom Id type.
Original pull request: #796.
Jens Schauder
Mark Paluch
Christoph Strobl
Denis Zavedeev
Nick Stolwijk