RFR+RFA [11u] (XS) 8219251: Langtools tests default memory size needs to be 768m
Thomas Stüfe
thomas.stuefe at gmail.com
Wed Feb 20 06:45:38 UTC 2019
Can live with the no. Issue was stuck in approvement limbo for ages anyway.
For the record, the reason why I wanted this was not "nice to have" - maybe
the text was too vague.
In very short, Runtime.exec uses vfork() on 11u, which may cause issues
which are extremely difficult to analyse, e.g. sporadic deaths of the
parent process, or trashing the forking threads stack. Since these problems
are so difficult to pinpoint it would have been nice to have an off switch
to try an alternate fork method in cases where we see rare random process
deaths.
..Thomas
On Tue, Feb 19, 2019 at 2:30 PM Langer, Christoph <christoph.langer at sap.com>
wrote:
> > Why is JDK-8212828 included in this list? It's not a bug. I can't see
> any strong
> > reason to include it. "Nice to have" is not really enough: there are
> lots of
> > things
> > that would be nice to have.
>
> We have to defer this one to Thomas. I think he started the backporting
> work some days ago when still Oracle was in charge. However, he needed to
> request a CSR for the backport - that's where this one is stuck at the
> moment. I agree, it should not be approved at the moment.
>
>
> > I ACK'd the rest.
>
> Thanks
>
More information about the jdk-updates-dev
mailing list