[Fwd: Re: Enhance footprint for array initialization]
John.Rose at Sun.COM
Thu Sep 4 10:47:43 PDT 2008
Hi, Ulf. I whitelisted you on the mailing list, since you are not
yet a member and it was blocking you.
You probably want to join, in order to receive any responses from the
JSR 292 is discussing bytecode modifications. Array init. is outside
the scope of JSR 292, so I don't know if we can do anything there in
the current round of changes.
But, at least, someone should prototype a mechanism as a subproject
of mlvm. That's what mlvm is for: It is an incubator for new JVM
On Sep 4, 2008, at 4:14 AM, Ulf Zibis wrote:
> Hi all,
> I had just opened a thread, about class file footprint, and the
> result was, that there is some need, to enhance it's format for
> array initialization using the constant pool:
> >From Patrick Wright I got the hint, to send you a copy of this
> discussion, because you would be "pretty open to hearing requests/
> ideas for enhancements to the JVM
> itself". ;-)
> ... so here I am. 8-)
> -------- Original-Nachricht --------
> Betreff: Re: Enhance footprint for array initialization
> Datum: Thu, 4 Sep 2008 12:31:28 +0200
> Von: Patrick Wright <pdoubleya at gmail.com>
> An: Ulf Zibis <Ulf.Zibis at gmx.de>
> Referenzen: <48BC243B.7050503 at gmx.de> <5B949BAF-FE81-4EC7-A772-
> C397E339E07D at sun.com> <48BE5ED5.2000308 at gmx.de>
> <1ccfd1c10809031912x23b8fe6dxbc193c686ae51aca at mail.gmail.com>
> <48BFB5D0.2080803 at gmx.de>
> Hi Ulf
> You might try re-posting on the DaVinci Machine mailing list. They are
> pretty open to hearing requests/ideas for enhancements to the JVM
> Currently they are focused on dynamic language support (the
> invoke-dynamic bytecode), but at the very least you'll probably get a
> detailed response from a senior HotSpot engineer about the
> Good luck!
> On Thu, Sep 4, 2008 at 12:17 PM, Ulf Zibis <Ulf.Zibis at gmx.de> wrote:
> > Martin, you bring it to the point in the end. Short and clear.
> > About starting an JSR... I wonder that not somebody else had
> tried this
> > in the past.
> > For me, I'm afraid, that I don't have the power for it. Thanks for
> > warning me.
> > Am 04.09.2008 04:12, Martin Buchholz schrieb:
> >> And discuss your efforts in this direction in core-libs-dev.
> > Yes. from this point on it belongs to core-libs-dev.
> > Ulf
> mlvm-dev mailing list
> mlvm-dev at openjdk.java.net
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the mlvm-dev