RFR: 8297168: Provide a bulk OopHandle release mechanism with the ServiceThread [v2]

[David Holmes]

On Mon, 21 Nov 2022 09:41:39 GMT, Robbin Ehn <rehn at openjdk.org> wrote:

>> Why? This is internally used code and any usage bugs should be exposed by testing. There is very little that can go wrong here.
>
> First to signal to a reader that the VM is utterly broken if this happens.
> Secondly when running test we also run on release, having the guarantee makes triage this failure to same as the assert much faster.
> 
> Keep assert if you want.

I prefer to keep the assert. I think guarantees should be reserved for things that are unexpected/highly-unlikely but still remotely possible due to external code actions - which is not the case here. Else every assert becomes a guarantee.

-------------

PR: https://git.openjdk.org/jdk/pull/11254