[NEW BUG]: Reduce allocations in sun.reflect.annotation.AnnotationInvocationHandler.invoke()

Claes Redestad claes.redestad at oracle.com
Mon Nov 28 14:20:53 UTC 2016


Hi,

doing a bit of digging it appears a similar improvement - along with a
lot of other things - was suggested and filed a few years back:

https://bugs.openjdk.java.net/browse/JDK-8029019

Peter, this enhancement is currently assigned to you, but I guess
you're not actively working on it. Any objection if I create a subtask
to deal with the observed performance issue in
AnnotationInvocationHandler.invoke - or do you want to revisit the
enhancement in full?

Thanks!

/Claes

On 2016-11-28 14:46, Christoph Dreis wrote:
> Hi,
>
>   
>
> Thank you for the quick response. I've signed the OCA already according to
> the contribution guidelines - I just wasn't sure where to put such a trivial
> enhancement.
>
>   
>
> In a 10 minute profile I had around 600MB of allocations for cloning the
> types, so roughly 1MB/s. If you need the screenshot, I can post it tomorrow.
>
>   
>
> Anything you need from me in order to proceed?
>
>   
>
> Cheers,
>
> Christoph
>
>   
>
>   
>



More information about the core-libs-dev mailing list