MAKE_JOBS_NUMBER to 1. This makes it safe to do -j${MAKE_JOBS_NUMBER}
without any extra logic.
- Cleanup ports working around the empty MAKE_JOBS_NUMBER
- This also fixes several ports that were expecting MAKE_JOBS_NUMBER
to always have a number
Reviewed by: bapt
Spotted by: John Marino <draco@marino.st>
With hat: portmgr
For a while now bsdtar is able to autotermine compression and archive format.
Let's then use tar directly instead of piping to tar.
Now USE_BZIP2 and USE_XZ only set EXTRACT_SUFX to the right format
http://icedtea.classpath.org/hg/release/icedtea6-1.11/rev/d9564350faa6http://blog.fuseyism.com/index.php/2012/10/19/security-icedtea-1-10-10-1-11-15-2-1-3-2-2-3-2-3-3-released/
- Completely turn off parallel build by default and remove parallel build
hack for HotSpot. There were several reports that it fails to build under
certain environment, ports/162991 for example. Users can still do parallel
build by setting FORCE_MAKE_JOBS (and MAKE_JOBS_NUMBER if desired).
- Implement os::available_memory(). Now it is consistent with "vm.vmtotal"
sysctl(3) MIB rather than bogus (physical memory / 4).
- Prefer sysconf(_SC_NPROCESSORS_CONF) over HW_NCPU sysctl MIB to get the
number of installed processors. There is no functional difference except
for CURRENT, which obtains the information from ELF aux vector.
- Prefer sysconf(_SC_PHYS_PAGES) * sysconf(_SC_PAGESIZE) over HW_USERMEM
sysctl MIB to get size of physical memory. Although it looks more logical
to find currently available memory, it has an inevitable side-effect, i. e.,
it changes dynamically depending on current wired page count. Therefore,
it is unpredictable and not too useful some times. For example, launcher
uses the parameter to determine initial heap size and machine class for i386.
Now it is more consistent with other places (and Linux JDK/JREs, including
the ones we have in ports tree).
- Implement os::active_processor_count() using cpuset_getaffinity(2). For
example, Runtime.getRuntime().availableProcessors() now returns number of
available processors for the current process as it should.
- Sync. launchers (java_md.c) for HotSpot and JDK as much as possible for
maintainability. As a good side-effect, launcher for i386 can now determine
machine class based on the current hardware configuration. Previously,
client VM was always chosen by default.
- Fix CounterGet(), which is only used for debugging launcher.
- Add swap info for os::print_memory_info().
Obtained from: IcedTea project [1]
Feature safe: yes
bootstrapped with pre-built openjdk6 packages from FreeBSD build clusters
on clean 7.4-RELEASE installation, and the ports tree was checked out at
r305682.
Feature safe: yes
manually set, this dependency may not be fulfilled.
- Consistently use WRKSRC over WRKDIR where it makes sense. Although they
are actually the same directory, there is no reason to confuse users.
Feature safe: yes
- Set JAVACMD environment variable to point to bootstrap JDK for ant. When
it is unset or empty, ant sets it by executing `which java`, which may be
different from bootstrap JDK.
the JDK build looks for them, but FreeBSD 9.x has them available in /usr/bin.
Adjusting the JDK check is a harder fix, this will get things building
again on 9.x quickly.
PR: 166996
result turn it on by default again. This change bears a little more
explanation.
In UnixFileSystem.java there is a cache of canonical paths and in
particular one for paths inside java.home. The problem with looking
up time zone data is that if another path within java.home has been
looked up then the symlink for the time zone data is not resolved
(since the cache assumes there are no symlinks in java.home) and we fail
to be able to open the ZoneInfoMapping file.
One of the other features in UnixFileSystem is that the cache is not
consulted and full resolution occurs for paths containing "." or "..".
I've chosen to (ab)use this to force resolution of the time zone data
directory in ZoneInfoFile.java rather than doing something like IcedTea
did which was to introduce a tz.properties file where they would stash
the time zone info directory rather than symlinking it in. For full
reference that patch is here:
http://icedtea.classpath.org/hg/icedtea6/file/tip/patches/use-system-tzdata.patch
There are two main reasons I chose a bit of a hack rather than something
like this.
1. I want to apply the patch to non-GPL'ed code (e.g. the jdk16 port).
2. Using updated time zone data is only an option in the port and using
something like tz.properties when the option is off is clunky.
Thanks to avg@ for sleuth work on finding the bug write up and the
IcedTea fix.
CVE-2011-3563: Fix issues in java sound
CVE-2011-3571: Fix in AtomicReferenceArray
CVE-2011-5035: Add property to limit number of request headers to the HTTP Server
CVE-2012-0501: Better input parameter checking in zip file processing
CVE-2012-0502: Issues with some KeyboardFocusManager method
CVE-2012-0503: Issues with TimeZone class
CVE-2012-0505: Enhance exception throwing mechanism in ObjectStreamClass
CVE-2012-0506: Issues with some method in corba
Obtained from: IcedTea Project
This is largely done for java detection for browser compatibility
Please update UPDATE_VERSION for new releases
PR: ports/163916
Submitted by: Denny Lin <dennylin93 at hs.ntnu.edu.tw>
Approved by: glewis, (rene, crees) (mentors,implicit)
- Import more complete JPEG class support and turn on ICEDTEA option by
default. [1]
- Remove obsolete WEB option. Note the functionality is now moved to a new
port java/icedtea-web, which depends on ICEDTEA option of this port.
Obtained from: IcedTea project [1]
off by default) and sort options. Add a basic wrapper to resurrect
com.sun.image.codec.jpeg using javax.imageio under the new option.
Obtained from: icedtea.classpath.org
PR: java/155288
is disabled by default because MIDI plugin for OSS does not exist, i.e.,
(hardware) MIDI sequencer is not supported. However, softsynth (Gervill)
works fine as it uses sampled sound.
virtual memory, free physical memory, and number of open file descriptors.
For example, jconsole shows more realistic numbers instead of hardcoded bogus
numbers in VM Summary tab now. Correct a jtreg test case for the swap info.
- Mention procfs(5) is used for some features in this implementation.
- Improve Serviceability Agent. For example, the following should work now:
jdb -connect sun.jvm.hotspot.jdi.SAPIDAttachingConnector:pid=<PID>
jmap <PID>
jsadebugd <PID>
jstack -F <PID>
jstack -m <PID>
- Replace IcedTea6 1.9.x with IcedTea-Web 1.0 for web support. Note it is
still tightly coupled with openjdk6 for now because we do not have proper
infrastructure to decouple them yet.
- Add itweb-settings (a Java Control Panel replacement for IcedTea-Web).
- Add desktop integration files for web support (but not installed in default
system-wide locations).
process. Update comments in the patch accordingly.
- Tidy up makefile for the plugin. Remove a hack for include path of libxul,
which is correctly patched from libxul 1.9.2.9_1. Remove mozilla-plugin.pc from
linker options for pkg-config. Generally speaking, NPAPI plugins do not need
to link with Gecko engine directly.
- Move run-time dependencies for the plugin to build-time dependencies as it is
completely optional any way.
Note the feature is turned off by default because it adds experimental and
unofficial patches from icedtea.classpath.org. You need to select WEB option
from config menu or define WITH_WEB variable to enable the feature. Also,
note that a symlink for web browser is not created automatically with same
reasoning. Thus, you need to create a symlink for your browser, e.g.,
ln -s ${LOCALBASE}/openjdk6/jre/lib/IcedTeadPlugin.so ~/.mozilla/plugins
Remove some stale comments in package message while I am here.
make jobs. If not, fall back to the number of CPUs, as per the current
behaviour.
Only the HotSpot part of the build is safe to parallelise, so leave the
port marked MAKE_JOBS_UNSAFE.
add IcedTea project as an additional master site. This change should fix
most build failures related to network problems, e.g., bootstrapping with
an IPv6-enabled JDK, unreliable network connections, etc.
- Add support for running regression tests
- Respect LOCALBASE
- Bump PORTREVISION
These are the test results on:
FreeBSD 6 i386: passed: 3113; failed: 29; error: 2
FreeBSD 8 amd64: passed: 3107; failed: 35; error: 2
Note: a couple of tests passed only after a few runs.
Submitted by: ale
Approved by: maintainer
- Fix font rendering and update cacerts file
- Add support for fastdebug jdk
- Cleanup/Enhance the port [2]
Submitted by: maintainer [1], ale [2]
Approved by: maintainer
Brian Gardner). Core support needs some more work.
. Install the debug build when WITH_DEBUG is set.
. Change WITH_DEBUG to only build the debug build.
. Eliminate the requirements for the binary plugins.
. Fix some compilation errors in the debug build on 7.x.
Submitted by: Kurt Miller <kurt@intricatesoftware.com>
was leaving Expert's Exchange, so I imagine email to him there is just
bouncing.
. Tweak the license description in COMMENT to a more standard form.
. Remove USE_JAVA. The JDK bootstrap dependency is handled separately in
the Makefile itself and once openjdk6 is part of bsd.java.mk USE_JAVA
could make this port depend on itself (which doesn't play nice with
portmaster).
. Host the patchset in the ports infrastructure rather than downloading it.
. Make the port build HotSpot rather than using the one of the bootstrapping
JDK (fixes a license violation).
Submitted by: Kurt Miller <kurt@intricatesoftware.com>
Approved by: Fixes a license violation
OpenJDK (6 or 7) doesn't yet come with a plugin or webstart.
Submitted by: Kurt Miller <lists@intricatesoftware.com>
http://docs.FreeBSD.org/cgi/mid.cgi?200902251003.14760.lists
Approved by: Brian Gardner <brian@experts-exchange.com> (maintainer)
