RFR: Workaround possible ResolvedMethodTable bug

[Roman Kennke]

Am 12.07.2017 um 20:39 schrieb Aleksey Shipilev:
> On 07/12/2017 08:32 PM, Roman Kennke wrote:
>> Hmmhmm, OK. But we need to understand and fix this soon.
> Yup. This would be the task for tomorrow.
>
>> The gotcha is that 'weak' doesn't always mean the same ;-)
> Please elaborate?

Just what you said:
>> "normal" weak references get processed fine, because they will get
registered with ReferenceProcessor, and will get marked if actually
reachable.

Then there are so-called 'weak roots' which have nothing to do with
java.lang.ref.Reference and related stuff. Strong GC roots always need
to be marked through, weak GC roots only under certain conditions (for
example, from the top of my head, if we don't want any class unloading,
we mark through all weak roots too).

Roman