On 12:27 Tue 18 Jan , Joe Darcy wrote: > On 1/11/2011 1:03 PM, Dr Andrew John Hughes wrote: > > On 11:36 Tue 11 Jan , Joe Darcy wrote: > >> On 1/11/2011 10:19 AM, Dr Andrew John Hughes wrote: > >>> On 09:43 Tue 11 Jan , Joe Darcy wrote: > >>>> Joe Darcy wrote: > >>>>> On 1/10/2011 12:24 PM, Dr Andrew John Hughes wrote: > >>>>>> On 11:26 Mon 10 Jan , Joe Darcy wrote: > >>>>>> > >>>>>> snip... > >>>>>> > >>>>>>> Hello. > >>>>>>> > >>>>>>> Since you haven't run into any issues with having 7002666 in IcedTea6 > >>>>>>> 1.9, I'm satisfied to have it this build of OpenJDK 6, which should > >>>>>>> bring b21 to a close! > >>>>>>> > >>>>>>> -Joe > >>>>>> I've pushed 7002666: > >>>>>> > >>>>>> http://hg.openjdk.java.net/jdk6/jdk6/hotspot/rev/499d2ae376ee > >>>>>> > >>>>>> along with two PNG backports that were approved for Deepak Bhole some > >>>>>> time > >>>>>> ago but were never done: > >>>>>> > >>>>>> http://hg.openjdk.java.net/jdk6/jdk6/jdk/rev/0e72e58a762d > >>>>>> http://hg.openjdk.java.net/jdk6/jdk6/jdk/rev/c267a33e76b0 > >>>>>> > >>>>>> (copyright headers updated appropriately). > >>>>>> > >>>>>> Let's get b21 out the door! :-) > >>>>> Agreed! Later this week, I'll go through my pre-promotion build and > >>>>> test cycle. > >>>>> > >>>>> -Joe > >>>> When running the HotSpot regression tests, I see two failures: > >>>> > >>>> Compilation failed: Compilation failed > >>>> > >>>> * gc/6581734/Test6581734.java : CMS Old Gen's collection usage is > >>>> zero after GC which is incorrect > >>>> > >>>> Execution failed: exit code 127 > >>>> > >>>> * runtime/6929067/Test6929067.sh : Stack guard pages should be > >>>> removed when thread is detached > >>>> > >>>> The first is due to a typo in the test case; I approve the backport of > >>>> 6980392 "TEST_BUG: gc/6581734/Test6581734.java has typo" > >>>> http://hg.openjdk.java.net/jdk7/tl/hotspot/rev/21c29458b334 > >>>> to fix this. > >>>> > >>>> The second test seems to make platform assumptions that aren't always > >>>> true, but after some looking around I don't see an existing fix in the > >>>> HotSpot repos. > >>>> > >>>> At least the first should be fixed before b21 promotes. > >>>> > >>> Yes, both of these issues are fixed in IcedTea6, the first with the > >>> 6980392 > >>> that I've now pushed to OpenJDK6. > >> Thanks for doing the push. > >> > >>> I recall there was some discussion of a fix for the second with upstream, > >>> but > >>> I can't find it right now and Pavel (who found the bug and fixed it) is on > >>> vacation. > >>> > >> One Oracle bug for this test failure seems to be, > >> > >> 6945460 TEST_BUG: runtime/6929067/Test6929067.sh > >> > >> but it is not marked as fixed in any release. > >> > >> There was some discussion on IcedTea: > >> > >> > >> http://mail.openjdk.java.net/pipermail/distro-pkg-dev/2010-November/011057.html > >> > >> and at least some on HotSpot: > >> > >> > >> http://mail.openjdk.java.net/pipermail/distro-pkg-dev/2010-November/010959.html > >> > >> http://mail.openjdk.java.net/pipermail/distro-pkg-dev/2010-December/011329.html > >> > >> but I don't see a resolution. > >> > > Yes, these are all on the IcedTea mailing list. I recall some discussion > > on the HotSpot mailing lists but I can't find any. > > > > The patch is in IcedTea6 as patches/jtreg-6929067-fix.patch. As I said, > > Pavel will > > know more but isn't around to comment at the moment. > > > > I'll take a took at the patch; I'd prefer b21 to have a clean HotSpot > regression test run and a fix for that issue will be necessary for that > to happen (I'll other HotSpot regression tests pass). > > -joe >
Ok I'll ping Pavel directly tomorrow. Let's not delay b21 much longer, it's already over six months since a release. There are unreleased javac fixes dating back to last June. -- Andrew :) Free Java Software Engineer Red Hat, Inc. (http://www.redhat.com) Support Free Java! Contribute to GNU Classpath and IcedTea http://www.gnu.org/software/classpath http://icedtea.classpath.org PGP Key: 94EFD9D8 (http://subkeys.pgp.net) Fingerprint = F8EF F1EA 401E 2E60 15FA 7927 142C 2591 94EF D9D8