-Note: I fail to reproduce the failure on 18.4.x/18.5.x+ (by running it until it fails on intellij, it's not failing after 1k+ executions) I do reproduce on 16.10.x/17.10.x Currently trying to investigate what changed, the tests themselves have not evolved much afaics.-
Update: I do reproduce on 18.x, it's just taking longer. The bug seems to be a real concurrency issue on ReadWriteSemaphore, but this is beyond my concurrency skills.
Manuel Leduc on 18/Jun/26 15:54
-Note: I fail to reproduce the failure on 18.4.x/18.5.x+ (by running it until it fails on intellij, it's not failing after 1k+ executions) I do reproduce on 16.10.x/17.10.x Currently trying to investigate what changed, the tests themselves have not evolved much afaics.-
Update: I do reproduce on 18.x, it's just taking longer. The bug seems to be a real concurrency issue on ReadWriteSemaphore, but this is beyond my concurrency skills. Possibly because lockWrite calls {{this.writeCounter.incrementAndGet()}} then {{this.semaphore.acquireUninterruptibly(this.readCounter.get() + 1)}} in a non-atomic way
This message was sent by Atlassian Jira (v9.3.0#930000-sha1:287aeb6)
If image attachments aren't displayed, see this article.
