June 2012 - core-libs-dev

hg: jdk8/tl/jdk: 7163471: Licensee source bundle failed around 7u4
by nils.loodin＠oracle.com 07 Jun '12

07 Jun '12

Changeset: c89018e3f3b6 Author: nloodin Date: 2012-06-05 13:43 +0200 URL: http://hg.openjdk.java.net/jdk8/tl/jdk/rev/c89018e3f3b6 7163471: Licensee source bundle failed around 7u4 Reviewed-by: dholmes, sla, brutisso, erikj ! make/com/oracle/Makefile

1 0

hg: jdk8/tl/jdk: 7164256: EnumMap clone doesn't clear the entrySet keeping a reference to the original Map
by alan.bateman＠oracle.com 07 Jun '12

07 Jun '12

Changeset: 757a5129fad7 Author: alanb Date: 2012-06-07 12:31 +0100 URL: http://hg.openjdk.java.net/jdk8/tl/jdk/rev/757a5129fad7 7164256: EnumMap clone doesn't clear the entrySet keeping a reference to the original Map Reviewed-by: alanb, chegar, forax, mduigou Contributed-by: dbelfer(a)gmail.com ! src/share/classes/java/util/EnumMap.java + test/java/util/EnumMap/ProperEntrySetOnClone.java

1 0

hg: jdk8/tl/jdk: 7175011: ProblemList.txt updates (6/2012)
by alan.bateman＠oracle.com 07 Jun '12

07 Jun '12

Changeset: 636f9486fde7 Author: alanb Date: 2012-06-07 10:31 +0100 URL: http://hg.openjdk.java.net/jdk8/tl/jdk/rev/636f9486fde7 7175011: ProblemList.txt updates (6/2012) Reviewed-by: mduigou, chegar ! test/ProblemList.txt

1 0

7175011: ProblemList.txt updates (6/2012)
by Alan Bateman 07 Jun '12

07 Jun '12

We have two test failures that look like they will take a while to get resolved in jdk8/tl. One of them is the FPU precision issue that Kelly brought up a few days ago. The other is a jconsole test that was missed by some recent refactoring in that area. I'd like to put these tests on the ProblemList.txt so that they are excluded from execution until they are fixed. Attached are the diffs. -Alan. diff --git a/test/ProblemList.txt b/test/ProblemList.txt --- a/test/ProblemList.txt +++ b/test/ProblemList.txt @@ -150,6 +150,9 @@ sun/management/jmxremote/startstop/JMXSt ############################################################################ # jdk_math + +# 7174532 +java/lang/Math/WorstCaseTests.java generic-all ############################################################################ @@ -374,6 +377,9 @@ com/sun/jdi/ProcessAttachTest.sh # Filed 6979016 sun/tools/jconsole/ResourceCheckTest.sh generic-all +# 7172176 +sun/tools/jconsole/ImmutableResourceTest.sh generic-all + # 7132203 sun/jvmstat/monitor/MonitoredVm/CR6672135.java generic-all

2 1

Review Request 7174736 : JCK test api/java_util/HashMap/index_EntrySet failing
by Mike Duigou 07 Jun '12

07 Jun '12

Hello all; Alan Bateman tracked down a regression introduced by the Alternative HashMap changes which was caught by the JCK tests. The fix is a one liner: diff --git a/src/share/classes/java/util/HashMap.java b/src/share/classes/java/util/HashMap.java --- a/src/share/classes/java/util/HashMap.java +++ b/src/share/classes/java/util/HashMap.java @@ -611,7 +611,7 @@ public class HashMap<K,V> Map.Entry<?,?> entry = (Map.Entry<?,?>) o; Object key = entry.getKey(); - int hash = (key == null) ? 0 : hash(key.hashCode()); + int hash = (key == null) ? 0 : hash(key); int i = indexFor(hash, table.length); @SuppressWarnings("unchecked") Entry<K,V> prev = (Entry<K,V>)table[i]; Which is unfortunately similar to another error discovered during pre-commit review. Barring complaints, I will push this change on Thursday. Mike

1 0

hg: jdk8/tl/jdk: 2 new changesets
by daniel.daugherty＠oracle.com 07 Jun '12

07 Jun '12

Changeset: 181175887d24 Author: jonas Date: 2012-06-06 13:16 -0700 URL: http://hg.openjdk.java.net/jdk8/tl/jdk/rev/181175887d24 7174861: all/OPT jdk build on Solaris with FDS enabled sets wrong options Summary: Use CFLAGS_COMMON instead of CC_OPT and CXXFLAGS_COMMON instead of CXX_OPT for setting FDS options. FDS should also set OPTIMIZATION_LEVEL. Reviewed-by: ihse, dholmes, ohair, dcubed ! make/common/Defs-solaris.gmk Changeset: 7543b6db3026 Author: dcubed Date: 2012-06-06 19:11 -0700 URL: http://hg.openjdk.java.net/jdk8/tl/jdk/rev/7543b6db3026 Merge

1 0

hg: jdk8/tl/jdk: 7172149: ArrayIndexOutOfBoundsException from Signature.verify
by xuelei.fan＠oracle.com 07 Jun '12

07 Jun '12

Changeset: 713b10821c3d Author: xuelei Date: 2012-06-06 18:39 -0700 URL: http://hg.openjdk.java.net/jdk8/tl/jdk/rev/713b10821c3d 7172149: ArrayIndexOutOfBoundsException from Signature.verify Summary: take care of integer addition overflow Reviewed-by: xuelei, wetmore Contributed-by: Jonathan Lu <luchsh(a)linux.vnet.ibm.com> ! src/share/classes/java/security/Signature.java + test/java/security/Signature/VerifyRangeCheckOverflow.java

1 0

hg: jdk8/tl/jdk: 7174244: NPE in Krb5ProxyImpl.getServerKeys()
by xuelei.fan＠oracle.com 07 Jun '12

07 Jun '12

Changeset: f8e72d7ff37d Author: xuelei Date: 2012-06-06 18:18 -0700 URL: http://hg.openjdk.java.net/jdk8/tl/jdk/rev/f8e72d7ff37d 7174244: NPE in Krb5ProxyImpl.getServerKeys() Reviewed-by: weijun ! src/share/classes/sun/security/ssl/SSLContextImpl.java ! src/share/classes/sun/security/ssl/ServerHandshaker.java ! src/share/classes/sun/security/ssl/krb5/Krb5ProxyImpl.java + test/sun/security/ssl/sanity/ciphersuites/CipherSuitesInOrder.java

1 0

An Alternative Hashing Alternative
by Doug Lea 06 Jun '12

06 Jun '12

I just posted the following to the concurrency-interest list. I'll send a follow-up on tie-ins to core-lib issues next. ... Finally acting on an old idea, I committed an update to ConcurrentHashMap (currently only the one in our jdk8 preview package, as jsr166e.ConcurrentHashMap8) that much more gracefully handles maps that are huge or have many keys with colliding hash codes. Internally, it uses tree-map-like structures to maintain bins containing more nodes than would be expected under ideal random key distributions over ideal numbers of bins. Among other things, this provides graceful (O log(N)) degradation when there are more than about a billion elements (which run out of 32bit hash resolution and array bound limits). However, the map can only do so if keys are Comparable, which is true of most keys in practice -- Strings, Longs, etc. Without relying on Comparable, we can't otherwise magically find any other means to further resolve and organize keys after running out of bits, so performance for non-Comparable keys is unaffected/unimproved. This also provides a mechanism for coping with hostile usages in which many keys (Strings in particular) are somehow constructed to have the same hashCode, which can lead to algorithmic denial of service attacks (because of linear-time bin searches) if code using the map don't screen external inputs. The overflow-tree-based strategy here is an alternative approach to adding secondary randomly-seeded hash code methods ("hash32") to class String, as has been committed recently to OpenJDK. But ConcurrentHashMap8 doesn't doesn't rely on this. The use of overflow-bins also allowed a few other minor speedups in more typical usages. A few more small improvements are still left unfinished for now. (I'll be traveling and/or otherwise committed for most of the next few weeks.) Links: jsr166e.jar: http://gee.cs.oswego.edu/dl/jsr166/dist/jsr166e.jar source: http://gee.cs.oswego.edu/cgi-bin/viewcvs.cgi/jsr166/src/jsr166e/ConcurrentH… Please give this a try and let me know about experiences. -Doug

3 7

hg: jdk8/tl/jdk: 7173645: (props) System.getProperty("os.name") should return "Windows Server 2012" for Windows Server 2012
by kurchi.subhra.hazra＠oracle.com 06 Jun '12

06 Jun '12

Changeset: af313ded4ffb Author: khazra Date: 2012-06-06 11:37 -0700 URL: http://hg.openjdk.java.net/jdk8/tl/jdk/rev/af313ded4ffb 7173645: (props) System.getProperty("os.name") should return "Windows Server 2012" for Windows Server 2012 Summary: Enable Windows Server 2012 to be recognized as "os.name" Reviewed-by: alanb, dholmes, chegar ! src/windows/native/java/lang/java_props_md.c

1 0