Changes include:
- any bug-fixes and code generator improvements
- new Thread.Abort implementation,
- I/O libraries
- speed improvements
... and much more
On >= NetBSD 2.0, if PERL5_USE_THREADS is _not_ defined, add pthread to libs.
This does not make perl threaded, but permits it to link against libraries
which are. This makes p5-GD (which pulls in threaded freetype) work again.
- fix a core dump in GC.
- fix pthread_np usage on NetBSD (we have the function, but not the
header pthread_np.h)
- documentation moved to share/doc
- OSX build fixes
- fix a GC lookup race conditon
through, unacceptably breaks threaded apps, and was made without consulting
the package's maintainer.
To quote my mail to packages@:
I am reversing this change.
With this change, tcl is _not_ usable in threaded apps, while without
this change, tcl is perfectly usable in non-threaded apps.
The fix for a package which is incorrectly written is not to break the
package it depends on, but do your job correctly when writing the
package.
With buildlink3, this is mind-numbingly easy to do.
That you neither contacted the maintainer of this package, nor did the
buildlink and dependent package changes which your change requires shows
that you have not really thought this through.
So, as I said, I am reversing this change. Fix your own package, don't
break others.
libtcl.so/libtk.so (whether linking it in or dlopen() it) must be linked
against libpthread. This is something we can't guarantee.
Fixes bulk build error in py-imagingtk and numerous runtime problems.
(restartable system calls). (These patches have been submitted back to the
python community.). Fixes the bug in NetBSD PR [pkg/24797] that I submitted.
* Highlight in the DESCR files what the thread support is for that package.
* Bump PKGREVISION.
Note that this cannot be changed to ${INSTALL} because it is not the install
command but rather an argument to setup.py. While adding the quotes makes
no difference to the use of the file, it does keep pkglint quiet. This is
even more important because this file is included by many other packages
and causes warnings when trying to check new Python extensions.
sun-jre13-1.0.12 and sun-jdk13-1.0.12 respectively. Patches provided by
Hauke Fath in PR pkg/25870 and PR pkg/25871. This update done during the
freeze as Sun removes the older distfiles when they release a new version.
Changes since latest version, according to
http://java.sun.com/j2se/1.3/ReleaseNotes.html:
- stack yellow zone not re-enabled after a StackOverflowError
- Gregorian calendar gives a java.lang.Illegal ArgumentException only on 12-31-00
- [server] stack yellow zone not re-enabled after a StackOverflowError
- compilation results in stack overflow while unrolling loop
- rmic uses AWT ?!
- -Xcheck:jni uses oop directly in _thread_in_native state
- Loop related Hotspot crash
- internal Error occurs during offet conversion of byte code in rewrite/relocate
- [1.3.1_09]JVM crashes when big number is specified in fillRect()
- Japanese characters not converting correctly from Codepage 930 to Codepage 943
- JVM crash with error "Fatal: null exception in compiled code"
- Client blocks on getInputStream even after server completes its response
- Error ID: e:\ws\1.3.1\hotspot\src\share \vm\ci\ciInstanceKlass.cpp, 121
- 5 JCK tests failing with -Xcheck:jni flag
by package Makefiles as well as buildlink3s to prevent usage of particular
versions of Python -- even if they appear in PYTHON_VERSIONS_ACCEPTED.
(And make Darwin exclusion use this variable instead, so that it effectively
overrides any pkg's definition of _ACCEPTED.)
This is a bug-fix release for Python 2.3 that fixes a number of bugs,
including a couple of weakref bugs and a bug in pickle version 2. There
are also a number of fixes to the standard library, and some build
fixes - see the release notes ( http://www.python.org/2.3.4/NEWS.html )
for details.
