Apache Ant 1.8.0 has been out now since February 8, 2010. The latest version in portage is still 1.7.1 An ebuild does not appear to be available in the Java overlay. Ant is available at http://ant.apache.org Reproducible: Always
Worked on it yesterday, got it mostly done locally. Let this bug be a tracker, to block with packages that break etc. From my brief run of re-emerge-all-java-packages there seems to be some that most probably started failing due to it: * ('ebuild', '/', 'dev-java/ant-contrib-1.0_beta3', 'merge'), Log * ('ebuild', '/', 'dev-java/jarjar-1.0', 'merge'), Log * ('ebuild', '/', 'dev-java/jgroups-2.7.0', 'merge'), Log * ('ebuild', '/', 'app-misc/freemind-0.9.0_rc6', 'merge'), Log
jarjar 1.0 (in java-experimental overlay) fixed, 0.9 in tree is fine
jgroups fixed with a patch in the 2.9.0 bump - will need to go stable before/with ant 1.8
freemind fixed with app-misc/freemind-0.9.0_rc7
dev-java/ant-contrib-1.0_beta3 fixed
Unmasked. Make any future related breakage block this bug.
Before someone complains -- the OpenJDK build inside IcedTea will fail with this new version of Ant. Don't yet know why.
1.8.1 added few days ago. I think it's better stable candidate than 1.8.0, which means we'll have to wait for a while. Which kind of sucks due to bug 303149.
Time to stabilize ant 1.8. The list of packages is below, should be the same for all arches. Since ant is used heavily for building many java packages through java-ant-2 eclass, I will be attaching lists of stable package-slots using the class. It would be nice if at least some arch could test this, although I suppose there should be no problems. Package Version Current Keywords Masks ============================= =================== ================= ========= dev-java/ant-swing 1.8.1 ~amd64 K dev-java/ant-testutil 1.8.1 ~amd64 K dev-java/ant-jsch 1.8.1 ~amd64 K dev-java/ant-jmf 1.8.1 ~amd64 K dev-java/ant-jdepend 1.8.1 ~amd64 K dev-java/ant-javamail 1.8.1 ~amd64 K dev-java/ant-jai 1.8.1 ~amd64 K dev-java/ant-commons-net 1.8.1 ~amd64 K dev-java/ant-commons-logging 1.8.1 ~amd64 K dev-java/ant-apache-resolver 1.8.1 ~amd64 K dev-java/ant-apache-regexp 1.8.1 ~amd64 K dev-java/ant-apache-oro 1.8.1 ~amd64 K dev-java/ant-apache-log4j 1.8.1 ~amd64 K dev-java/ant-apache-bsf 1.8.1 ~amd64 K dev-java/ant-apache-bcel 1.8.1 ~amd64 K dev-java/ant-antlr 1.8.1 ~amd64 K dev-java/ant-apache-xalan2 1.8.1 ~amd64 K dev-java/ant-trax 1.8.1 ~amd64 K dev-java/ant-junit 1.8.1 ~amd64 K dev-java/ant-nodeps 1.8.1 ~amd64 K dev-java/ant-core 1.8.1 ~amd64 K dev-java/ant 1.8.1 ~amd64 K
Created attachment 235409 [details] x86 list of stable package-slots using java-ant-2 eclass
Created attachment 235411 [details] amd64 list of stable package-slots using java-ant-2 eclass
Created attachment 235413 [details] ant-ppc.txt =ppc list of stable package-slots using java-ant-2 eclass
Created attachment 235415 [details] ant-ppc64.txt ppc64 list of stable package-slots using java-ant-2 eclass
I tested almost everything on x86 (so far i only took one version (latest stable) per package, if there were more than one in slots). It seems that you forgot dev-java/ant-junit4-1.8.1 in the list for stabilizing, so i also tested that one with its rdeps. I only found a few thigs: dev-util/netbeans-5.5.1-r1 will not build with it! The stable-Bug #318151 could help, but also has some nasty blockers. net-im/openfire - every version in portage claims "Must use Ant 1.6.x or 1.7.x to build Openfire". app-emacs/jde-2.3.5.1-r2 doesn't compile currently, but 2.4.0.1 is since feb in the tree without bugs and it passed all tests with the newer ant and all rdeps seem to be ok with it. dev-java/metadata-extractor-2.2.2-r2 as well as the unstable metadata-extractor-2.2.2-r3 do not compile! It work's with ant 1.7. dev-java/piccolo2d-1.2.1-r1 is the latest version in the tree and also doesn't compile. This also works with ant 1.7!
(In reply to comment #14) > I tested almost everything on x86 (so far i only took one version (latest > stable) per package, if there were more than one in slots). > It seems that you forgot dev-java/ant-junit4-1.8.1 in the list for stabilizing, > so i also tested that one with its rdeps. Thank you for those extensive tests. > dev-util/netbeans-5.5.1-r1 will not build with it! The stable-Bug #318151 could > help, but also has some nasty blockers. > > net-im/openfire - every version in portage claims "Must use Ant 1.6.x or 1.7.x > to build Openfire". > > app-emacs/jde-2.3.5.1-r2 doesn't compile currently, but 2.4.0.1 is since feb in > the tree without bugs and it passed all tests with the newer ant and all rdeps > seem to be ok with it. Emacs team will ask for stabilisation soon, this should be a non-issue then. > dev-java/metadata-extractor-2.2.2-r2 as well as the unstable > metadata-extractor-2.2.2-r3 do not compile! It work's with ant 1.7. > > dev-java/piccolo2d-1.2.1-r1 is the latest version in the tree and also doesn't > compile. This also works with ant 1.7! > Everything else: Java team can we have a thumbs up for the other packages or a transition plan?
(In reply to comment #14) Thanks for the testing! > I tested almost everything on x86 (so far i only took one version (latest > stable) per package, if there were more than one in slots). > It seems that you forgot dev-java/ant-junit4-1.8.1 in the list for stabilizing, > so i also tested that one with its rdeps. Right, thanks, please add: dev-java/ant-junit4-1.8.1 > net-im/openfire - every version in portage claims "Must use Ant 1.6.x or 1.7.x > to build Openfire". Fixed. > I only found a few thigs: > dev-util/netbeans-5.5.1-r1 will not build with it! The stable-Bug #318151 could > help, but also has some nasty blockers. > > dev-java/metadata-extractor-2.2.2-r2 as well as the unstable > metadata-extractor-2.2.2-r3 do not compile! It work's with ant 1.7. > > dev-java/piccolo2d-1.2.1-r1 is the latest version in the tree and also doesn't > compile. This also works with ant 1.7! Cannot reproduce any Ant specific errors here. However I've noticed they fail with icedtea6(-bin) and work with sun-jdk-1.6. For piccolo2d it's bug 322663, metadata-extractor bug 297999 (none for netbeans yet). So I assume your ant 1.7 environment had sun-jdk and 1.8 had icedtea6(-bin) which led you to this conclusion? > Everything else: Java team can we have a thumbs up for the other packages or a > transition plan? Assuming I'm correct and you stabilize the new emacs jde mentioned, there are no more blockers :) (and we shouldn't need new netbeans in bug 318151 neither)
amd64 stable
meh, mid-air collision
(In reply to comment #16) > Cannot reproduce any Ant specific errors here. However I've noticed they fail > with icedtea6(-bin) and work with sun-jdk-1.6. For piccolo2d it's bug 322663, > metadata-extractor bug 297999 (none for netbeans yet). So I assume your ant 1.7 > environment had sun-jdk and 1.8 had icedtea6(-bin) which led you to this > conclusion? > Yes, you were right with piccolo2d and metadata-extractor... Is there a way to force a specific java-env!? I tried with java-config -S, but only an unmerge solved it! But i couldn't get netbeans to compile... jar-prep: [mkdir] Created dir: /var/tmp/portage/dev-util/netbeans-5.5.1-r1/work/netbeans-src/nbbuild/netbeans/platform6/lib jar: [jarwithmoduleattributes] Building jar: /var/tmp/portage/dev-util/netbeans-5.5.1-r1/work/netbeans-src/nbbuild/netbeans/platform6/lib/org-openide-util.jar # # A fatal error has been detected by the Java Runtime Environment: # # Internal Error (nmethod.cpp:1836), pid=11290, tid=3066772336 # Error: guarantee(cont_offset != 0,"unhandled implicit exception in compiled code") # # JRE version: 6.0_20-b02 # Java VM: Java HotSpot(TM) Client VM (16.3-b01 mixed mode, sharing linux-x86 ) # An error report file with more information is saved as: # /var/tmp/portage/dev-util/netbeans-5.5.1-r1/work/netbeans-src/nbbuild/hs_err_pid11290.log # # If you would like to submit a bug report, please visit: # http://java.sun.com/webapps/bugreport/crash.jsp # /var/tmp/portage/dev-util/netbeans-5.5.1-r1/temp/environment: line 945: 11290 Aborted ant ${antflags} "${@}"
(In reply to comment #19) > (In reply to comment #16) > Yes, you were right with piccolo2d and metadata-extractor... Is there a way to > force a specific java-env!? I tried with java-config -S, but only an unmerge > solved it! Either edit jdk.conf according to http://www.gentoo.org/doc/en/java.xml#preferred-vm or after new java-config stabilized in bug 321767, java-config -S will work as well. > But i couldn't get netbeans to compile... > > # Internal Error (nmethod.cpp:1836), pid=11290, tid=3066772336 > # Error: guarantee(cont_offset != 0,"unhandled implicit exception in compiled > code") > # > # JRE version: 6.0_20-b02 > # Java VM: Java HotSpot(TM) Client VM (16.3-b01 mixed mode, sharing linux-x86 ) > # An error report file with more information is saved as: Definitely JVM error, not ant.
(In reply to comment #20) > Definitely JVM error, not ant. > Nope... I just downgraded the ant packages and netbeans builds again!!
(In reply to comment #21) > (In reply to comment #20) > > > Definitely JVM error, not ant. > > > Nope... I just downgraded the ant packages and netbeans builds again!! Still JVM error, just triggered by some specific version of ant. No java code can make JVM crash unless there's bug in JVM. Please file a different bug saying which jdk that was and attaching the error log file. It should not block ant stabilization anyway.
stable x86, thanks Andreas
Created attachment 250737 [details] new ppc list, with versions On ranger's request, giving new ppc list with exact versions instead of slots.
Created attachment 250739 [details] new ppc64 list, with versions same for ppc64
ppc64 done
ppc, can you do it? I doubt much testing is needed at this point
ppc stable, sorry that it took so long