----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/52805/ -----------------------------------------------------------
Review request for geode, Barry Oglesby and nabarun nag. Repository: geode Description ------- When executing a function from a client, we can be adding results to the result collector from multiple threads. Our docs claim the user should not have to synchronize their result collector. One code path was already synchronizing on the collector when adding results. However, if the function returned an exception we were not synchronizing. Adding a SynchronizedResultCollector and wrapping the users collector in that to ensure that there will be no unsynchronized access of the result collector. Diffs ----- geode-core/src/main/java/org/apache/geode/cache/client/internal/ExecuteFunctionOp.java 6597b680227fb47492dc973baecffd504d6cdf68 geode-core/src/main/java/org/apache/geode/cache/client/internal/ExecuteRegionFunctionSingleHopOp.java 51ea8e4fbc64acbbd1165856a2dd09704d63e857 geode-core/src/main/java/org/apache/geode/internal/cache/execute/ServerFunctionExecutor.java 4295898583aff1409bf6acdd84c5a4c8a8709e51 geode-core/src/main/java/org/apache/geode/internal/cache/execute/ServerRegionFunctionExecutor.java b5bc684e0b29bfc62dd958ccff49d47c2bad36fa geode-core/src/main/java/org/apache/geode/internal/cache/execute/util/SynchronizedResultCollector.java PRE-CREATION Diff: https://reviews.apache.org/r/52805/diff/ Testing ------- Ran the affected test 500 times with this fix. It failed 11 times without it, passed with the fix. Thanks, Dan Smith
