Upstream code has seen many structural changes and the failing test is gone.
Since changing to latest sources might cause huge reworking/test-efforts this
patch only fixes the build error.
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Fix stripping related warnings
Drop foldarith_for_spe patch not neded anymore
QA Error about elocations in .text are still there like 2.0.0
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
According to make.tmpl.in, shared lib will be installed under $(usrlibdir)
install_lib_shared: $(LIB_SHARED)
$(INSTALL_PROGRAM) -D $< $(libdir)/$(<F).$(LIB_VERSION)
$(INSTALL_DIR) $(usrlibdir)
$(LN_S) -f $(USRLIB_RELPATH)$(<F).$(LIB_VERSION) $(usrlibdir)/$(<F)
But default $(usrlibdir) is $(DESTDIR)${prefix}/lib no matter it is 32-bit or 64-bit system.
So we need align $(usrlibdir) with $(libdir).
Signed-off-by: Olivia Yin <hong-hua.yin@freescale.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
WARNING: QA Issue: yajl: Files/directories were installed but not shipped
/usr/lib
/usr/lib/libyajl.so.2
/usr/lib/libyajl.so
/usr/lib/libyajl_s.a
/usr/lib/libyajl.so.2.0.4
/usr/lib/.debug
/usr/lib/.debug/libyajl.so.2.0.4
NOTE: for 64-bit systems where lib64 is used you can pass the cmake
variable LIB_SUFFIX to cause installation into the system's 'lib64'
directory.
Signed-off-by: Olivia Yin <hong-hua.yin@freescale.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Use the headers from the linux-libc-headers instead of the headers
from the linux kernel sources because user space programs should
be built against the header files provided by the distribution.
In file included from .../usr/src/kernel/include/linux/media.h:27:0,
from conftest.c:56:
.../usr/src/kernel/include/linux/types.h:13:2: warning: #warning \
"Attempt to use kernel headers from user space, see http://kernelnewbies.org/KernelHeaders" [-Wcpp]
On some platforms (qemu) intern kernel header (linux/types.h) redefines
some types from the system header (sys/types.h, timer.h, stdint.h)
and thereby breaks the checking of the linux media header.
ERROR: Kernel header file not found or not usable!
In file included from .../usr/include/linux/media.h:27:0,
from conftest.c:56:
In file included from .../usr/src/kernel/include/linux/types.h:14:26: error: conflicting types for 'fd_set'
In file included from .../sys/select.h:75:5: note: previous declaration of 'fd_set' was here
Similar errors for dev_t, mode_t, timer_t, uintptr_t and blkcnt_t.
Signed-off-by: Stefan Herbrechtsmeier <stefan@herbrechtsmeier.net>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
PTYS_ARE_GETPT and PTYS_ARE_SEARCHED don't appear anywhere in the
source code anymore, remove them.
Signed-off-by: Jack Mitchell <jack.mitchell@dbbroadcast.co.uk>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
* now I see why it was there before, at least on arm it fails with:
| CMake Error at cmake/modules/LLVMConfig.cmake:127 (message):
| Library `X86' not found in list of llvm libraries.
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
* when building for qemux86* it tries to build X86 twice and fails with:
| CMake Error at lib/Target/CMakeLists.txt:24 (add_subdirectory):
| The binary directory
|
| /home/jenkins/oe/shr-core-branches/shr-core/tmp-eglibc/work/x86_64-oe-linux/llvm2.9/2.9-r2.0/llvm-2.9/build/lib/Target/X86
|
| is already used to build a source directory. It cannot be used to build
| source directory
|
| /home/jenkins/oe/shr-core-branches/shr-core/tmp-eglibc/work/x86_64-oe-linux/llvm2.9/2.9-r2.0/llvm-2.9/lib/Target/X86
|
| Specify a unique binary directory name.
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Exiting postinst code leaves systemd.bbclass offline postinst code unexecuted,
causes postinst not to finish at first boot (lxdm.service is not aliased to
display-manager.service -> started much too early) and is not neccesary:
language preparation can be done offline.
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
* Change DESCRIPTION to SUMMARY
* Fix LICENSE declarations to be version-specific
* Coding convention fixes * Don't modify Makefiles modified by qmake2; instead
symlink qmake2 in to $WORKDIR to be run from there so that it uses
$WORKDIR/qt.conf values
Signed-off-by: Wayne Warren <steven.w.warren@gmail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Based on initial version by Koen Kooi <koen@beagleboard.org>. Initscript
borrowed from Debian with some tweaks.
Note that since collectd uses libltdl, building this successfully
required OE-Core commit db84eaf851b22b262d9dc48eb55bd5224a00fdd2 or else
you get an error about "config/compile" being missing.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
* Set SUMMARY instead of DESCRIPTION
* Move SRC_URI checksums under SRC_URI
* Move packaging definitions to the end
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
* llvm is generic enough to be useful outside meta-java
e.g. we need it to enable llvmpipe galium driver in mesa
* imported without any modification (except indentation), all
credits go to Henning Heinold and Khem Raj for maintaining
it in meta-java
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
The whitespace in LICENSE causes bitbake to parse this as three separate
licenses: 'Arphic', 'Public' and 'License'. These spaces are replaced with
dashes to prevent confusion.
Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Replace version 7.2.446 with the latest version currently available, 7.3.905.
The old fetch method was inefficient as version 7.2 was downloaded and patches
applied to get to version 7.2.446. Attempting this method with git 7.3 did not
work as not all 905 patches applied cleanly. Instead we fetch version 7.3.905
directly from the upstream mercurial repository. This causes an implicit
dependency on mercurial-native for the build which is reasonable considering
that not many images include vim by default.
The previously used patch configure.in_remove_CC_quotes.patch is still required
as these changes haven't been made upstream; this patch still applies cleanly.
Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
