Path to this page:
Subject: CVS commit: pkgsrc/lang
From: Jan Schaumann
Date: 2003-10-29 23:18:54
Message id: 20031029221854.B81A92DA1D@cvs.netbsd.org
Log Message:
Trivially update sun-j*14 to the latest version (sun-j*14-2.2, corresponding
to Sun's 1.4.2_02). This is done even though we are in a freeze since
older distfiles disappear and a security vulnerability in older versions
might make people update.
Changes/bugfixes since 1.4.2_01, according to
http://java.sun.com/j2se/1.4.2/ReleaseNotes.html
- Regression: IllegalStateException:constrain(xywh) not supp. for complex transform
- AsyncGetCallTrace() on Solaris/Intel core dumps with assertion failure
- occasional crashes/deadlocks from AsyncGetCallTrace()
- JVMPI_EVENT_COMPILED_ METHOD_LOAD: "must release all locks when leaving \
VM" asser
- Cannot set preferences factory in web start
- crash in AsyncGetCallTrace() when invoked during Deoptimization
- AsyncGetCallTrace() is absent in linux/Intel JVM
- infinite recursion in AsyncGetCallTrace() with the server JVM
- syncGetCallTrace: assert(pc_desc != 0, "scopeDesc must exist").
- AsyncGetCallTrace: guarantee(get_thread() == thread, "must be the same thread.
- JNI GetEnv() race with ThreadLocalStorage::set_thread() guarantee
- Cache versioning fails to work in 1.4.1 and higher releases
- Get/SetThreadLocalStorage need to work with AsyncGetCallTrace
- JInternalFrame setDefaultCloseOperation ignored under Motif look-and-feel.
- OBJ_ALLOC event deadlock between VMThread and Finalizer thread
- VM crash when HeapDump requested
- Util buffer cache uses wrong release strategy
- Need general-purpose way to dispatch object invocations on Java GUI thread
- Crash in OopFlow::build_oop_map
- [mercury] Java Setup menuitem location misleading
- Update cacerts file with GeoTrust root certs
- extra pings to server if WeakReferenced CachedJarLoaders gc'ed
- Size of sliders in GTK L&F larger than original
- Open/Closing buttons on split panes are very small
Files: