RFR: 8250913: EventHandlerCreator embeds an unnecessary StringPool
Erik Gahlin
erik.gahlin at oracle.com
Thu Aug 6 14:00:22 UTC 2020
Hi Chihiro,
Plan was to use a string pool local to the field, but we never got around to it when we did the clean-room implementation for JDK 9. We had this in Oracle JDK 8. For example, it is possible to cache the value of a field and if it is the same as the next, it could only then be promoted to the global pool. Coming up with a heuristics requires benchmarking, which takes time.
We could remove the implementation and dig it up in the version control system if we want to do the optimization later, so I’m fine with the fix. That said, if we are going eliminate all traces of field local caching, more changes are needed and in my opinion not worth spending time on, especially since we may add it back.
Thanks
Erik
> On 2 Aug 2020, at 17:40, Chihiro Ito <chiroito107 at gmail.com> wrote:
>
> Hi,
>
> Could you review this fix, please?
>
> This fix passed submit-repo.
>
> JBS : https://bugs.openjdk.java.net/browse/JDK-8250913
> Webrev : http://cr.openjdk.java.net/~cito/JDK-8250913/webrev.00/
>
> Regards,
> Chihiro
More information about the hotspot-jfr-dev
mailing list