| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
| |
IcedTea 2.6.2 provides OpenJDK 7u91, which adds the October 2015 security
updates. For more details see:
http://blog.fuseyism.com/index.php/2015/10/23/security-icedtea-2-6-2-for-openjdk-7-released/
Furthermore the patches directory for OpenJDK-7 was renamed to avoid the version number.
This makes it easier for upgrades and to see diff's between patches.
Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Hotspot compiler has a bug where signed integers could overflow. This is
undefined behaviour, and causes massive memory leak when compiled with GCC 5.0+,
causing the build to fail.
This is fixed by backporting patch from
https://bugs.openjdk.java.net/browse/JDK-8078666 where it was fixed by
Severin Gehwolf.
Signed-off-by: Erkka Kääriä <erkka.kaaria@intel.com>
Signed-off-by: Maxin B. John <maxin.john@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Upstream-Status: Backport
Fixes undefined behaviour in boehm GC used by cacao-initial-native
as suggested at
https://github.com/ivmai/bdwgc/commit/16780405c899f01119bc45167122b4242f6b34fb.
The undefined behaviour at very least causes a segmentation fault during
classpath-native configure step with Fedora 22.
Signed-off-by: Erkka Kääriä <erkka.kaaria@intel.com>
Signed-off-by: Maxin B. John <maxin.john@intel.com>
|
|
|
|
|
|
|
|
|
| |
Since Shark is utterly broken and cacao and jamvm are terrible slow compared
to zero, it's pointless to support them for targets with limited resources
unless there is active maintainership taken by responsible person.
Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
| |
Since Yocto Project has a reasonable definition of supported build hosts, we
shouldn't restrict to e.g. pre 4.x kernels as icedtea7 expects.
Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
| |
Since OpenJDK-6 is not maintained anymore, it's more reasonable to move to OpenJDK-8 instead
of sticking at unmaintained, upstream not supported VM version.
Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I realized one small issue last days: the dependency of openjdk-*-demos on java2-vm, which is
provided by the current recipe (eg. openjdk-7) causes bitbake scan for alternatives (eg. java2-vm
provided by openjdk-6) without any need.
I would recommend to remove such a generic dependency:
* openjdk-6-demos require openjdk-6-jre, period.
* openjdk-7-demos require openjdk-7-jre, period.
* openjdk-8-demos require openjdk-8-jre, period.
There is no sense in satisfying openjdk-8-demos with openjdk-6-jre (even if it might be possible).
Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
| |
Allow building icedtea7 without any X11 dependency.
This affects the native part only which is required to bootstrap openjdk-8-native
for cross-compiling openjdk-8, but should be adoptable for cross-compiled recipe
as well as for openjdk-6.
Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The recipe uses sed to patch src/scripts/java.in so that it execs cacao-initial
instead of cacao. However, if configure step is run again, -initial is appended
again to the string, leading to 'exec cacao-initial-initial'. This causes the
build to fail.
This is fixed by changing the sed command to consider larger portion of the
source string. This ensures that the -initial is only appended once.
Signed-off-by: Erkka Kääriä <erkka.kaaria@intel.com>
Signed-off-by: Maxin B. John <maxin.john@intel.com>
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit 15ed9662e26ed4ac8d843fb9d2335a49a9c03088.
The change has break external builds causing build failures. The
change here needs to be revisited to avoid the breakages and than
re-applied.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
BitBake raises some errors when processing recipes that depend on x11
for distros that don't have x11 in DISTRO_FEATURES. To work around
that issue, REQUIRED_DISTRO_FEATURES = "x11" (from
distro_features_check.bbclass) has been set for the following recipes:
* cacao
* classpath
* jamvm
* openjdk-6-release
* openjdk-7-release
That makes BitBake skip those recipes during the cache generation
(they'd still be parsed, but ignored).
Signed-off-by: Mario Domenech Goulart <mario@ossystems.com.br>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The recipes below depended on a hardcoded native compile
(ecj-initial-native). Since we may use another native compiler to
build them, we set the dependency in virtual/javac-native instead.
* cacao
* classpath-native
* jamvm
Signed-off-by: Mario Domenech Goulart <mario@ossystems.com.br>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
| |
OpenJDK 7 was previously updated to IcedTea 2.6.1, 7u85b01. Therefore
the old version is removed now.
Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
| |
Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
| |
Increment the PR of openjdk-7-common.inc due to the previous changes.
Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
| |
Use the PN variable instead of the hardcoded "openjdk-7-jre" string.
Furthermore fix the indentation of its content.
Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
| |
The "--disable-pulse-java" configuration is no longer available in
IcedTea 2.6.1, therefore move it to the 2.5.4 (75b13) version only.
Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
| |
PCSC default value was changed to true and SCTP was added during the 2.6.0
IcedTea release so disable them both explicitely in openjdk-7-common.inc
Signed-off-by: Richard Leitner <richard.leitner@skidata.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Rather than just use d.getVar(X), use the more explict d.getVar(X, False)
since at some point in the future, having the default of expansion would
be nice. This is the first step towards that.
This patch was mostly made using the command:
sed -e 's:\(getVar([^,()]*\)\s*):\1, False):g' -i `grep -ril getVar *`
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
| |
Fixing what looks like a hand edited patch error
Signed-off-by: Amy Fong <amy.fong@windriver.com>
Signed-off-by: Maxin B. John <maxin.john@enea.com>
|
|
|
|
|
|
|
| |
We have now updated to IcedTea 2.5.4, so we can remove the old version.
Signed-off-by: Sven Ebenfeld <sven.ebenfeld@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
| |
Signed-off-by: Sven Ebenfeld <sven.ebenfeld@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
| |
Due to problems when using native rmic, I switched to using icedtea7.
Signed-off-by: Sven Ebenfeld <sven.ebenfeld@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
| |
Signed-off-by: Sven Ebenfeld <sven.ebenfeld@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
| |
Signed-off-by: Sven Ebenfeld <sven.ebenfeld@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
| |
Signed-off-by: Sven Ebenfeld <sven.ebenfeld@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
| |
Signed-off-by: Sven Ebenfeld <sven.ebenfeld@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
| |
The build system looks for xsltproc during configure, when bootstrap
is required, so we need to have it around. This has been catch when
building the recipe in a clean tmp directory.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
| |
Cacao uses an internal version of bdwgc that was being installed
(headers, libs, etc.) together with the JVM, unnecessarily conflicting
with original bdwgc package.
A patch from Gentoo is applied so the system's bdwgc is used instead.
Signed-off-by: Hugo Vasconcelos Saldanha <hugo.saldanha@aker.com.br>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
| |
Fixes:
WARNING: QA Issue: icedtea6-native: configure was passed unrecognised
options: --disable-npplugin
Signed-off-by: Mario Domenech Goulart <mario@ossystems.com.br>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
| |
Some recipes install JNI-based bindings into /usr/lib/jni (e.g.,
rxtx), so it's sensible to add this directory to DEFAULT_LIBPATH
(java.library.path). That's also how Debian does.
Signed-off-by: Mario Domenech Goulart <mario@ossystems.com.br>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
| |
Since bd9c6023, virtual/java-initial is provided by
cacao-initial-native and jamvm-initial-native. For consistency, this
commit makes them provide virtual/java-initial-native instead.
Signed-off-by: Mario Domenech Goulart <mario@ossystems.com.br>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
-jre binaries such as keytool, policytool and more requires libjli.so
so ensure there's a RDEPEND that provides it.
Same goes for openjdk-7-java (java binary FILERDEPENDS libjli.so).
Both of these already DEPENDS -common, so the build was not broken before
but this patch will also silence the QA file-rdeps issue.
Signed-off-by: Magnus Olsson <magnus@minimum.se>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A number of recipes in meta-java use autotools.bbclass for building; internally
it uses the sstate-control populate_sysroot manifests to locate and copy
aslocals for configure dependencies as part of the do_configure step.
The manifest path differs depending on if it's a native package or not so
autotools.bbclass looks at the package name to determine if its native or not
(it's native if it ends with "-native").
The warnings are emitted because a few (native) recipes in meta-java (such as
cacao-initial, classpath-initial and more) are incorrectly being classified as
non-native by autotools.bbclass, which makes it look for the manifest in the
wrong place and finally gives up with the warning:
WARNING: /home/oe/tmp-glibc/sstate-control/manifest-mymachine-libecj-bootstrap.populate_sysroot not found
This happens for cacao-initial, classpath-initial, ecj-initial, libecj-boostrap,
jamvm-initial and jikes-initial since they are all native and rely on autotools.
This patch renames the recipes so they end with -native, e.g ecj-initial-native.
It also updates the recipes so they DEPEND on the corresponding new name. This
helps autotools.bbclass to properly classify the recipes are native and silence
the warnings.
Signed-off-by: Magnus Olsson <magnus@minimum.se>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
| |
Removes a currency date expiration check whis causes a build
failure.
Signed-off-by: Hugo Vasconcelos Saldanha <hugo.saldanha@aker.com.br>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
| |
do_install was calling an unnecessary 'make install'. oe_jarinstall
provided by java-library class is sufficient.
Signed-off-by: Hugo Vasconcelos Saldanha <hugo.saldanha@aker.com.br>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
| |
In native build, the link is not created by update-alternatives.
Some package builds were failing, like gnumail and gnujaf, if cacao
was the PREFERRED_PROVIDER_virtual/java-native.
Something similiar is currently done with jamvm-native.
Signed-off-by: Hugo Vasconcelos Saldanha <hugo.saldanha@aker.com.br>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
| |
Before configuring package, we upgrade m4 macros with a call to
libtoolize. Otherwise, compiling would fail with a complain about
different versions of libtool.
Signed-off-by: Hugo Vasconcelos Saldanha <hugo.saldanha@aker.com.br>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
While testing commit 'java-library: inherit allarch' I found some recipes
throwing errors in do_compile.
The recipes jaxme and xom currently don't build with compiler errors.
jaxen, dom4j, jdom, commons-jxpath, commons-configuration depend directly
or indirectly on jaxme and/or xom and thus are unbildable.
Signed-off-by: Max Krummenacher <max.oss.09@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
| |
The override virtclass_native has been deprecated and removed from the
buildsystem, use class-native instead.
While at it fix whitespace issues.
Signed-off-by: Max Krummenacher <max.oss.09@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Use the allarch class to make every packages in the 'all' ARCH.
Recipes inheriting java-library which build additional packages which
don't belong in to all ARCH must manually setting PACKAGE_ARCH after
inheritting java-library.
This fixes the following error when building e.g. rhino for a second
machine of different architecture in the same build directory:
ERROR: The recipe rhino is trying to install files into a
shared area when those files already exist. Those files and their
manifest location are:
/.../ipk/all/librhino-java_1.7r4-r0_all.ipk
Matched in manifest-armv7at2hf-vfp-rhino-package_write_ipk
Signed-off-by: Max Krummenacher <max.oss.09@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
| |
While at it change the LICENSE string to the oe one if applicable.
e.g. AL-2.0 -> Apache-2.0
Signed-off-by: Max Krummenacher <max.oss.09@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
| |
Indent with 4 spaces in task definitions.
Signed-off-by: Max Krummenacher <max.oss.09@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is only way for compile for now since qemu can not
run the sizers binary for e6500 machines. It segfaults
so we disable generating sizes and also xawt and jawt
which depends on sizers, this lets jre to cross compile
for ppc/e6500
The fix is ported according to da0b3a20a67496dc8d4282d19a60672001b7f825
Signed-off-by: Zhenhua Luo <zhenhua.luo@freescale.com>
Signed-off-by: Ting Liu <ting.liu@freescale.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
set CROSS_COMPILE_ARCH which will be checked before run gamma test.
Fix the below issue:
| make[6]: Leaving directory `.../openjdk-6-jre/6b32-1.13.4-r22.0/build/openjdk.build/hotspot/outputdir/linux_ppc64_zero/product'
| cd linux_ppc64_zero/product && ./test_gamma
| ./test_gamma: line 72: ./gamma: cannot execute binary file
| make[5]: *** [productzero] Error 126
Signed-off-by: Ting Liu <ting.liu@freescale.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix the below issue:
| checking whether to enable the use of LCMS2... yes
| checking whether to use the system LCMS install... yes
| checking for LCMS2... no
| configure: error: Could not find LCMS >= 2.5; install it or build with --disable-system-lcms to use the in-tree copy.
| Configure failed. The contents of all config.log files follows to aid debugging
| ERROR: oe_runconf failed
Signed-off-by: Ting Liu <ting.liu@freescale.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As the openjdk source was extracted during 'make patch-ecj' in
do_configure, apply the patch via do_configure_append.
icedtea6-1.13.4/Makefile.am:1200:stamps/extract-openjdk.stamp: stamps/download.stamp
| if OPENJDK_SRC_DIR_FOUND
| if ! test -d openjdk ; then \
| cp -pPR $(OPENJDK_SRC_DIR_LINK) $(OPENJDK_SRC_DIR) openjdk ; \
| fi
| else
| if USE_HG
| if ! test -d openjdk ; then \
| cp -pPRl openjdk.hg openjdk ; \
| fi
| else
| set -e ; \
| if ! test -d openjdk ; \
| then \
| mkdir openjdk ; \
| $(TAR) xf $(OPENJDK_SRC_ZIP) -C openjdk; \
| chmod -R ug+w openjdk ; \
| sh $(abs_top_srcdir)/fsg.sh ; \
| fi
| endif
| endif
Signed-off-by: Ting Liu <ting.liu@freescale.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit ab51e51d9a9a51bedd19de09bcba4ce830dfc5e7.
Using an oe setup with ptest enabled makes qemu never to return due to testfiles
staged by udev in the sysroot
See also:
http://lists.openembedded.org/pipermail/openembedded-core/2014-September/097098.html
All binaries compiled for qemu are now linked with -static. The -L, -E parameters
pointing to the sysroot and dynamic linker are not needed anymore and removel works
around the issue.
Further details:
http://lists.openembedded.org/pipermail/openembedded-devel/2015-January/099795.html
Signed-off-by: Max Krummenacher <max.oss.09@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For machines using DEFAULTTUNE = "armv7ahf" or DEFAULTTUNE = "armv7athf" a build
of openjdk-7 fails with segmentation faults during the qemu-arm runs.
e.g.
Segmentation fault (core dumped) | qemu-arm -cpu cortex-a8 -s 2097152
-L .../sysroots/colibri-t20 \
-E LD_LIBRARY_PATH=.../sysroots/colibri-t20/lib ./mkbc - bytecodes_arm.s
I found that qemu-arm can not execute any binary which is compiled with the
resulting compiler and is linked with the gold linker and the -static flag.
Even a simple hello world program does not work. When executing that
hello world on the real hardware it also generates a segmentation fault.
Removing the -static flag or forcing the bfd linker fixed this.
Note that other machines, colibri-t30 or i.mx6 based ones with neon but
otherwise identical setup do not suffer from the issue.
DEFAULTTUNE = "armv7athf-neon" works.
While at it convert the sizer call for openjdk-7 used when building xawt
also to use -static even though the code is currently disabled.
Further details:
http://lists.openembedded.org/pipermail/openembedded-devel/2015-January/099795.html
Signed-off-by: Max Krummenacher <max.oss.09@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The currency data generation class makes a 10 years expiration date check.
It might be reasonable to make sure the input data is updated
periodically, but on an embedded system we don't want to have this type of
built in time expirations.
This commit removes the date expiration check.
Signed-off-by: Alex Gonzalez <alex.gonzalez@digi.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|