From: Mingli Yu
Remove _BUILD_STRING and _BUILD_TIME in _pl_bld.h
to avoid introducing timestamp in the .so library
such as libnspr4.so as below:
$ readelf --wide --decompress --hex-dump=.rodata libnspr4.so
[snip]
0x4000 32303138 2d31312d 31352030 353a3439 2018-11-15 05:49
[snip]
[YOCTO
On 2018-11-15 9:15 p.m., Paul Eggleton wrote:
Hi Bruce
I just came across this the other day and I had one question:
On Sunday, 19 August 2018 2:50:44 PM NZDT Bruce Ashfield wrote:
The existing kernel-devsrc package starts with a full copy of the kernel
source and then starts to strip out elem
From: Chunrong Guo
(From OE-Core rev: c67bc9bdcbcf46ed54dfbafd1bb6e405ed77f86a)
Signed-off-by: Chunrong Guo
---
meta/conf/machine/include/tune-ppce500v2.inc | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/meta/conf/machine/include/tune-ppce500v2.inc
b/meta/conf/machine/inc
Hi Bruce
I just came across this the other day and I had one question:
On Sunday, 19 August 2018 2:50:44 PM NZDT Bruce Ashfield wrote:
> The existing kernel-devsrc package starts with a full copy of the kernel
> source and then starts to strip out elements that are not required.
>
> This results
On Thu, 2018-11-15 at 16:28 -0600, Joshua Watt wrote:
>
> classes/testsdk(ext)?-oecore.bbclass
>
> python do_testsdk() {
> from oeqa.sdk.testsdk import TestSDK
> t = TestSDK(d)
> t.run()
> }
>
>
>
> It might make things a little more clear if I describe the end game;
On Thu, 2018-11-15 at 19:40 +, Richard Purdie wrote:
> On Thu, 2018-11-15 at 13:14 -0600, Joshua Watt wrote:
> > Splits the SDK test implementation into separate class files. The
> > classes that are included are controlled by the variable
> > ${TESTSDK_CLASSES}. This allows a SDK machine to ch
On Thu, 2018-11-15 at 10:02 +0800, Chen Qi wrote:
> bunzip2 is enabled by default, but bzip2 is not. This is kind of
> strange, and it also causes busybox's ptest failure regarding
> bunzip2, as bunzip2's test case needs bzip2 command.
>
> Signed-off-by: Chen Qi
> ---
> meta/recipes-core/busybox
On Thu, 2018-11-15 at 13:14 -0600, Joshua Watt wrote:
> Splits the SDK test implementation into separate class files. The
> classes that are included are controlled by the variable
> ${TESTSDK_CLASSES}. This allows a SDK machine to change which classes
> are used to implement the test cases.
>
> [
Splits the SDK test implementation into separate class files. The
classes that are included are controlled by the variable
${TESTSDK_CLASSES}. This allows a SDK machine to change which classes
are used to implement the test cases.
[YOCTO #13020]
Signed-off-by: Joshua Watt
---
meta/classes/tests
On Thu, 2018-11-15 at 12:00 +, Mike Crowe wrote:
> When using usrmerge, the compatibility dynamic loader symlink created
> in do_install_append_aarch64 ends up in /usr/lib rather than /lib.
> This means that the path added to libc_baselibs needs to be altered
> to match.
>
> It's also necessar
On Thu, 15 Nov 2018 at 01:57, Chen Qi wrote:
> Use example.org to ensure it's always reachible.
Is this because of China's firewall? You'd think that google.com
would have excellent uptime.
Ross
--
___
Openembedded-core mailing list
Openembedded-core
Two recent patches are now marked as reported in the
valgrind bugzilla.
Signed-off-by: Randy MacLeod
---
.../0001-Guard-against-__GLIBC_PREREQ-for-musl-libc.patch | 2 +-
...1-Make-local-functions-static-to-avoid-assembler-error.patch | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-
On Thu, Nov 15, 2018 at 9:35 AM Martin Jansa wrote:
> yes fw-utils is similar, but has different dependencies and builds slightly
> differently.
fw-utils use the machine setup to find the default environment to use,
so it is indeed machine specific.
--
Otavio Salvador
On Thu, 2018-11-15 at 15:00 +, Burton, Ross wrote:
> On Thu, 15 Nov 2018 at 14:50, Marek Vasut wrote:
> > I have to admit, the recipe name becomes a misnomer then. But
> > renaming
> > this recipe would be hard, since it's used all over the place. Any
> > ideas?
>
> Call it u-boot-tools, pote
On 11/15/2018 04:00 PM, Burton, Ross wrote:
> On Thu, 15 Nov 2018 at 14:50, Marek Vasut wrote:
>> I have to admit, the recipe name becomes a misnomer then. But renaming
>> this recipe would be hard, since it's used all over the place. Any ideas?
>
> Call it u-boot-tools, potentially merge the oth
On Thu, 15 Nov 2018 at 14:50, Marek Vasut wrote:
> I have to admit, the recipe name becomes a misnomer then. But renaming
> this recipe would be hard, since it's used all over the place. Any ideas?
Call it u-boot-tools, potentially merge the other recipe tool, and add
RPROVIDES for compatibility?
On 11/15/2018 01:43 PM, Alexey Brodkin wrote:
> This utility is used for creation of images containing
> usable in run-time U-Boot environment.
>
> Note it is already gets built with "make tools" command
> together with "mkimage" so we just need to install "mkenvimage".
>
> As of today this utili
996247ba7dfffbeb444f793f7e105fcfb5ffa939 moved the power
flags (POWEROFF, HALT, REBOOT) to a fragment init.cfg which
is only included when busybox is selected as the init_manager.
This assumption is incorrect in scenarios where installation
mechanism is provided through live images and almost all t
On Tue, 23 Oct 2018 at 01:38, Anibal Limon wrote:
>> > -fstypes = [fs for fs in d.getVar('IMAGE_FSTYPES').split(' ')
>> > +fstypes = d.getVar('IMAGE_FSTYPES')
>> > +fstypes = [fs for fs in fstypes if fs in supported_fstypes]
That's the problem, the old code takes IMAGE_FSTYPES an
Python 3.7 removed the os.errno import, so fix dnf to use the correct import.
Signed-off-by: Ross Burton
---
meta/recipes-devtools/dnf/dnf/errno.patch | 36 +++
meta/recipes-devtools/dnf/dnf_2.7.5.bb| 1 +
2 files changed, 37 insertions(+)
create mode 100644 met
On Thu, 15 Nov 2018 at 12:35, Burton, Ross wrote:
> This appears to break if musl is being used:
>
> In file included from ../../../git/lib/ext2fs/ext2fs.h:97,
> from ../../../git/lib/e2p/feature.c:20:
> ../../../git/lib/ext2fs/hashmap.h:20:19: error: missing binary
> operator bef
This utility is used for creation of images containing
usable in run-time U-Boot environment.
Note it is already gets built with "make tools" command
together with "mkimage" so we just need to install "mkenvimage".
As of today this utility is added per-board like here [1]
for Intel Edison board.
On Wed, 14 Nov 2018 at 10:47, Burton, Ross wrote:
> Getting reports that this breaks cross builds, for example python3-dbus.
>
> tmp-glibc/work/cortexa7hf-neon-vfpv4-oe-linux-gnueabi/python3-dbus/1.2.6-r0/recipe-sysroot-native/usr/lib/libpython3.7m.so:
> file not recognized: File format not recog
This appears to break if musl is being used:
In file included from ../../../git/lib/ext2fs/ext2fs.h:97,
from ../../../git/lib/e2p/feature.c:20:
../../../git/lib/ext2fs/hashmap.h:20:19: error: missing binary
operator before token "("
#if __GNUC_PREREQ (4, 8)
^
.
When using usrmerge, the compatibility dynamic loader symlink created in
do_install_append_aarch64 ends up in /usr/lib rather than /lib. This means
that the path added to libc_baselibs needs to be altered to match.
It's also necessary to tweak do_poststash_install_cleanup to not fail when
the syml
Fix the warnings:
meta/lib/oeqa/utils/qemurunner.py:250: DeprecationWarning: invalid escape
sequence \.
ips = re.findall("((?:[0-9]{1,3}\.){3}[0-9]{1,3})", cmdline.split("ip=")[1])
meta/lib/oeqa/utils/qemurunner.py:343: DeprecationWarning: invalid escape
sequence \-
if re.search("root@[a-zA-
list_pkgs() for rpm calls RpmPM() which would try and create a copy of the
package feed. This can be called for example from buildhistory whilst some
other task may be working on an SDK or image construction, causing the package
feed to disappear part way through an operation.
Avoid the need to co
Avoid the warning:
meta/lib/oeqa/selftest/context.py:8: DeprecationWarning: the imp module is
deprecated in favour of importlib; see the module's documentation for
alternative uses
import imp
In this case importlib is a direct replacement.
Signed-off-by: Richard Purdie
---
meta/lib/oeqa/se
Hi,
yes fw-utils is similar, but has different dependencies and builds slightly
differently.
mkenvimage on the other hand is already built by current u-boot-mkimage
recipe, just not installed by default (adding separate recipe would build
it twice).
On Thu, Nov 15, 2018 at 11:32 AM Alexey Brodki
On 15/11/18 06:05, Alexey Brodkin wrote:
> Hi Khem,
>
> On Wed, 2018-11-14 at 20:57 -0800, Khem Raj wrote:
>> On Wed, Nov 14, 2018 at 8:00 PM Alexey Brodkin
>> wrote:
>>> This utility is used for creation of images containing
>>> usable in run-time U-Boot environment.
>>>
>>> As of today this uti
Hi Martin,
On Thu, 2018-11-15 at 10:27 +0100, Martin Jansa wrote:
> On Thu, Nov 15, 2018 at 05:15:03AM +, Alexey Brodkin wrote:
> > Hi Marek,
[snip]
> > > Reviewed-by: Marek Vasut
> > >
> > > That said, could the recipe for mkimage and mkenvimage be somehow
> > > deduplicated ? There seems
On 11/15/2018 05:23 PM, Richard Purdie wrote:
On Thu, 2018-11-15 at 10:02 +0800, Chen Qi wrote:
busybox's echo test cases needs 'od' command to have '-t' option
support. However, our busybox disable it by default by not setting
CONFIG_DESKTOP. CONFIG_DESKTOP affects many busybox utilities. So
ad
On Thu, Nov 15, 2018 at 05:15:03AM +, Alexey Brodkin wrote:
> Hi Marek,
>
> On Thu, 2018-11-15 at 06:03 +0100, Marek Vasut wrote:
> > On 11/15/2018 05:57 AM, Khem Raj wrote:
> > > On Wed, Nov 14, 2018 at 8:00 PM Alexey Brodkin
> > > wrote:
> > > > This utility is used for creation of images c
On Thu, 2018-11-15 at 10:02 +0800, Chen Qi wrote:
> busybox's echo test cases needs 'od' command to have '-t' option
> support. However, our busybox disable it by default by not setting
> CONFIG_DESKTOP. CONFIG_DESKTOP affects many busybox utilities. So
> add coreutils to RDEPENDS of busybox-ptest.
The openssl build system generates buildinf.h containing the full
compiler command line used to compile objects. This breaks
reproducibility, as the compile command is baked into libcrypto, where
it is used when running `openssl version -f`.
Add stripped build variables for the compiler and cflags
The busybox.link.* files are generated from autoconf.h and applets.h,
which are both auto-generated by the build system. The contents of the
two files might be in different order, and so the link files are not
reproducble as is.
Fix this by sorting the lists using `sort`.
Signed-off-by: Martin Hu
The shadow configure script tries really hard to detect the running
shell to make sure it doesn't do unsupported calls.
On my system the shell is detected as /bin/sh, while a build in an
ubuntu docker it resolves to /bin/bash. And since the shell path is
baked into the target binaries through conf
Because libnfsidmap has been remove, enable libnfsidmap to keep consistency of
nfs-utils.
Signed-off-by: Lei Maohui
---
meta/recipes-connectivity/nfs-utils/nfs-utils_2.3.1.bb | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/meta/recipes-connectivity/nfs-utils/nfs-utils_2.
From: Changqing Li
yocto-check-layer will find locked-sigs.inc under builddir,
but locked-sigs.inc is generated under current bitbake working
dir. if run yocto-check-layer outside builddir, we will met error
like "No such file or directory: *locked-sigs.inc". change to
run bitbake -S under buildd
39 matches
Mail list logo