BewareMyPower commented on PR #24457: URL: https://github.com/apache/pulsar/pull/24457#issuecomment-3003628757
You can check [benchmark.diff.txt](https://github.com/user-attachments/files/20896048/benchmark.diff.txt) for how I set up the tests. BTW, @codelipenghui mentioned "over engineering" [here](https://github.com/apache/pulsar/pull/24457#discussion_r2165507421). However, IMO, the cache itself is "over engineering" to me. I checked many places that `TopicName` is used. Mostly it's just used as a temporary variable to do some conversion. Both the existing solution and this solution have issues when a lot of different topic names are used. The existing solution relies on manual call of `clearIfReachedMaxCapacity`. Introducing a scheduled task also brings unnecessary complexity. The solution in this PR relies on GC to clean up entries from cache. If any issue happens, it could be harder to diagnose. Actually, I think we should remove the cache, which brings many issues that should never happen. From my perspective, introducing the cache is just to get around the poor implementation of `TopicName`'s constructor and it's a "GC PTSD" coding style of Pulsar. -- 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: commits-unsubscr...@pulsar.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
