HotSpot 17 merge
erik.trimble at oracle.com
Sun May 16 16:40:51 PDT 2010
Martin Buchholz wrote:
> On Sun, May 16, 2010 at 09:14, Andrew John Hughes <ahughes at redhat.com> wrote:
>> The webrev is here:
> I'm a lttle worried about the stability of the resulting hotspot.
> It appears that the most stable hotspots are obtained by
> getting the hotspot versions that are targetting proprietary
> 6uN releases, with selected later bug fixes added on top.
> Recently, this means even-numbered hotspots.
> So I was expecting that we would upgrade openjdk6
> to hotspot18 sometime within the next 6 months or so,
> skipping hotspot17 entirely.
> Perhaps someone from hotspot-land could give us
> more guidance.
I'll have to look it up when I get in on monday (I'm OOTO for the weekend).
Escape Analysis was turned off in HS17 because it wasn't quite 100%
working (there was a couple of edge cases that it had regressions on),
and we don't have time to make it work in HS17.
The SSE4.2 stuff was contributed code that's broken, and we didn't
find/figure out until recently. It's being fixed, but the safe thing to
do is just turn it off for now. I likewise don't expect it to be fixed
in HS17 (though, I could be wrong).
HS17 should be the JVM intended for 6u22 - I have to look to see,
though. It definitely is being prepped as the next "stable release"
train for the 6uN series. And, I think I need to push the build tag
onto those repos.
Like I said, I'll look into it more on Monday.
Java System Support
Santa Clara, CA
Timezone: US/Pacific (GMT-0800)
More information about the jdk6-dev