OpenJDK 6 B25 Available - backported patches by the icedtea team

[Andrew Hughes]

----- Original Message -----
> 2012-05-11 22:17, Dalibor Topic skrev:
> > On 5/11/12 12:41 PM, Andrew Hughes wrote:
> >> I'll second that.  I wasn't even aware of this release until
> >> Deepak pointed
> >> it out last week.
> > 
> > Lana started the promotion process in April:
> > http://mail.openjdk.java.net/pipermail/jdk6-dev/2012-April/002666.html
> > 
> >> Are we still working on OpenJDK6?  The last time I posted
> >> a patch it was suggested that things were winding down, so I'm a
> >> little surprised
> >> to see another release.  We have a whole stack of patches still
> >> needing upstreaming
> >> if they are wanted.
> > 
> > I think that releases for OpenJDK 6 make sense for the CPUs until
> > November. Xerxes
> > mentioned some backports here:
> > http://mail.openjdk.java.net/pipermail/jdk6-dev/2012-April/date.html
> > - are those the
> > fixes you mean?
> 
> No those two new patches are still not merged into IcedTea and under
> bug triage,
> I brought them up for discussion to hopefully solve the two most
> frequent crashing bugs reported in to the IcedTea team.
> 

Why are they not at least in IcedTea?

> The following two icedtea bugs indicate something odd in OpenJDK 6
> synchronizer.cpp that look similar to the
> http://mail.openjdk.java.net/pipermail/jdk6-dev/2012-April/002669.html
> - JVM crashes when heavyweight monitors are used
> > #  Internal Error (synchronizer.cpp:1401), pid=1232, tid=1455453296
> > #  guarantee(mid->header()->is_neutral()) failed: invariant
> http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=887 - The bug
> for minecraft server on linux
> http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=904 - unhandled
> signal 11 : os_linux_zero.cpp:236
> 
> And the following bugs indicate something odd in OpenJDK 6 that looks
> similar to
> http://mail.openjdk.java.net/pipermail/jdk6-dev/2012-April/002670.html
> - arraycopy stubs place prebarriers incorrectly
> http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=874
> 
> I think Andrew Huges was refering to the following large set of
> backports currently in use in IcedTea 6 builds:
> These 100+ patches are all backports from the OpenJDK 7/8 hg trees to
> OpenJDK 6.
> http://icedtea.classpath.org/hg/icedtea6/file/6ae19889a9e0/patches/openjdk
> It would be really nice if all of these patches would be applied to
> OpenJDK 6 directly.

Yes, that's what I was referring to.  Over a hundred straight backports, plus
some other patches that need to take the full 8->7->6 route.

Sorry for the delayed reply.  It turns out I wasn't filtering jdk6-dev specifically
and it got lost in all the other OpenJDK traffic.

> 
> The IcedTea project also host a large set of new patches for OpenJDK
> 6 that are in various state of upstreaming progress:
> http://icedtea.classpath.org/wiki/IcedTea_JDK6_Patches
> When we manage to upstream one of these new patches then we quickly
> add a new backport to the IcedTea6 patch folders,
> patches/openjdk , based on the upstream reviewed OpenJDK 7 and 8
> commit.
> 
> Cheers
> Xerxes
> 
> 

-- 
Andrew :)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

PGP Key: 248BDC07 (https://keys.indymedia.org/)
Fingerprint = EC5A 1F5E C0AD 1D15 8F1F  8F91 3B96 A578 248B DC07