westonpace commented on PR #14334: URL: https://github.com/apache/arrow/pull/14334#issuecomment-1270807888
This approach would cause all future callbacks to be serialized which would change the dynamics of the tests a little bit. I currently rely on SleepABitAsync's thread-per-sleep behavior to ensure parallelism in my stress tests. So this might reduce the coverage of the stress tests. Can we use a ThreadPool instead? #14339 should be sufficient. This would mean we still have sufficient parallelism in the tasks. The only downside is that if you have something like 128 sleeps then you will end up waiting 4 * delta seconds (where delta is SleepABit). However, I did some experiments and this doesn't seem to be a problem. Plus, I think running the tasks in parallel is likely to make up for any time lost. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
