Burton, Ross ross.bur...@intel.com writes:
I ran into this problem the other day, and after speaking to Ross
Burton, he suggested that the better way to fix this is with:
RDEPENDS_${PN}-dev =
I'm starting to get the feeling that the dev-pkgs feature is also a
little broken, as it seems as
Robert Yang liezhi.y...@windriver.com writes:
possible bashism in autofs/etc/init.d/autofs line 116 ($foo should be
eval_gettext foo):
+RDEPENDS_${PN} += bash
Is this bashism only in the initscript, or somewhere else too? If it is
only in the initscript, the RDEPENDS should be a
The 'systemctl disable' operation should not be done in postrm() because
the .service file does not exist anymore and the script will fail hence.
After moving this command into prerm(), postrm() becomes empty and can
be removed completely.
Signed-off-by: Enrico Scholz enrico.sch...@sigma
Otavio Salvador ota...@ossystems.com.br writes:
This patch caches the result of 'git rev-list' into the sources cache
directory and uses it on the next run. Because collisions of the sha1
hash are very unlikely, the git revision is used directly as the key.
...
I'd prefer to have the hashes
-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
meta-oe/classes/gitpkgv.bbclass | 30 --
1 file changed, 24 insertions(+), 6 deletions(-)
diff --git a/meta-oe/classes/gitpkgv.bbclass b/meta-oe/classes/gitpkgv.bbclass
index 165ba1e..7bdc538 100644
--- a/meta-oe
unless the corresponding collection has been selected. As it
relies on the order of BBFILES and BBFILE_COLLECTIONS expansion, a
sanity check has been added which aborts the build when assumptions do
not apply anymore.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
meta-systemd
Samuel Stirtzel s.stirtzel-gM/ye1e23mwn+bqq9rb...@public.gmane.org
writes:
+ExecStart=/sbin/dhclient -cf /etc/dhcp/dhclient.conf -q $INTERFACES -lf
/var/lib/dhcp/dhclient.leases
...
systemd[1]: PID file /var/run/dhclient.pid not readable (yet?) after start.
[FAILED] Failed to start Dynamic
Patch allows to specify additional parameters (e.g. '-B') by reading
/etc/default/dropbear as a systemd environment file. It applies
$DROPBEAR_EXTRA_ARGS which used already by the oe-core sysv initscript.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
meta-systemd/meta-oe
not sure whether the custom do_compile_prepend() is still needed.
For now only the issue above will be fixed by executing ./config.status
yet again.
[1] see 648290d5bf4d6ff50d3643bb7ad902dfc23aa702
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
meta/recipes-devtools/libtool/libtool
Phil Blundell ph...@gnu.org writes:
Changing the architecture to x86-64 would be a possibility,
no; will cause too much pain because all the ./config.guess scripts
expect 'x86_64'.
I kept TARGET_ARCH='x86_64' but added an additional OVERRIDES where '_'
is replaced by '-'.
Enrico
Paul Menzel paulepan...@users.sourceforge.net writes:
I ran a build from scratch with `EXTRA_AUTORECONF = --exclude=autopoint`
from `autotools.bbclass` and did not see any regressions.
There are regressions in projects using 'intltool' (e.g. 'pmount');
./configure will fail with
Depending on autotools/libtool version and configure flags, dbus-daemon
will be built sometimes below .libs/ and sometimes in base directory.
Using 'libtool --mode=install' is a portable way to deal with this.
Changing Makefile.am to use native automake methods to install this
files would be a
The old dangling link detection used os.stat() on the file and
registered it as a dangling link when it was not found. This causes
problems when the link is an absolute one because it will check
files of the host system.
E.g. when there is a link
.../tmp/work/../etc/cron.root -
previously.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
classes/packaged-staging.bbclass | 23 ++-
1 files changed, 18 insertions(+), 5 deletions(-)
diff --git a/classes/packaged-staging.bbclass b/classes/packaged-staging.bbclass
index 9c72d11..f3648d2 100644
Enrico Scholz enrico.sch...@sigma-chemnitz.de writes:
Depending on autotools/libtool version and configure flags, dbus-daemon
will be built sometimes below .libs/ and sometimes in base directory.
Using 'libtool --mode=install' is a portable way to deal with this.
Changing Makefile.am to use
Enrico Scholz enrico.sch...@sigma-chemnitz.de writes:
The old dangling link detection used os.stat() on the file and
registered it as a dangling link when it was not found. This causes
problems when the link is an absolute one because it will check
files of the host system
Sebastian Spaeth sebast...@sspaeth.de writes:
Changing Makefile.am to use native automake methods to install this
files would be a better but more invasive fix.
Looks reasonable, although I am always for going with the better fixes
in general :-)
I will work on a better patch which will be
Koen Kooi koen.k...@gmail.com writes:
if [ -e include/asm ] ; then
# This link is generated only in kernel before 2.6.33-rc1,
don't stage it for newer kernels
ASMDIR=`readlink include/asm`
- mkdir -p ${STAGING_KERNEL_DIR}/include/$ASMDIR
-
Old find statement failed with something like
$ find . \( -name XXX\)
find: invalid expression; I was expecting to find a ')' somewhere but did not
see one.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
classes/packaged-staging.bbclass |2 +-
1 files changed, 1
:
| oe_runconf --with-build-cflags=${BUILD_CFLAGS}
This patch might create incompatibilities for recipes which workaround
the
current limitations.
As a sideeffect, 'set -x' debug output is now much nicer because
superflous whitespaces in the configure call are removed.
Signed-off-by: Enrico Scholz
Chris Larson clar...@kergoth.com writes:
This patch allows to pass quoted strings in EXTRA_OECONF.
I fail to see how this is a problem. Seems to work fine for every
other app, including make.
EXTRA_OEMAKE handling is implemented as
| ${MAKE} ${EXTRA_OEMAKE} $@
but EXTRA_OECONF as
|
Chris Larson clar...@kergoth.com writes:
Ahh, I see, that makes it clear, thanks for clarifying. I have a local
patch that turns cfgcmd into a bitbake variable, and runs ${cfgcmd}
$@.
What do you think about that approach as an alternative?
It will work... I do not know the story behind
Problems mentioned in 4187d3df8838506f6578a2b0f2fb98f98c41e151 (Aug
2006) seem to went away and build succeeds with system zlib now. Using
system libraries is preferable because it saves space and allows better
tracking of security issues.
Signed-off-by: Enrico Scholz enrico.sch...@sigma
From: Enrico Scholz enrico.sch...@informatik.tu-chemnitz.de
Bytecode stuff is patent encumbered so it can be enabled for
non-enterprise distributions only.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
.../freetype-2.3.12/bytecode-interpreter.patch | 22
That's (nearly) a copy of the old 2.3.11 recipe.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
.../freetype/freetype-2.3.12/fix-configure.patch | 13 ++
recipes/freetype/freetype-2.3.12/libtool-tag.patch | 20 ++
recipes/freetype/freetype_2.3.12.bb
pieterg piet...@gmx.com writes:
Bytecode stuff is patent encumbered so it can be enabled for
non-enterprise distributions only.
The bytecode patents have expired,
afaik (I am not a lawer) that's only true for the second part
(TT_CONFIG_OPTION_BYTECODE_INTERPRETER) of the patch. The
This patch replaces a complicated if statement with a more simple
code.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
classes/patch.bbclass | 12 +---
1 files changed, 5 insertions(+), 7 deletions(-)
diff --git a/classes/patch.bbclass b/classes/patch.bbclass
index
it is not a patch or the parameter is false.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
classes/base.bbclass |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/classes/base.bbclass b/classes/base.bbclass
index 21b5646..f21d8ad 100644
--- a/classes/base.bbclass
Problems mentioned in 4187d3df8838506f6578a2b0f2fb98f98c41e151 (Aug
2006) seem to went away and build succeeds with system zlib now. Using
system libraries is preferable because it saves space and allows better
tracking of security issues.
Signed-off-by: Enrico Scholz enrico.sch...@sigma
From: Enrico Scholz enrico.sch...@informatik.tu-chemnitz.de
Bytecode stuff is patent encumbered so it can be enabled for
non-enterprise distributions only.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
.../freetype-2.3.12/bytecode-interpreter.patch | 22
That's (nearly) a copy of the old 2.3.11 recipe.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
.../freetype/freetype-2.3.12/fix-configure.patch | 13 ++
recipes/freetype/freetype-2.3.12/libtool-tag.patch | 20 ++
recipes/freetype/freetype_2.3.12.bb
Raffaele Recalcati lamiapost...@gmail.com writes:
I was searching a clean way to develop in OE my new kernel and my new
applications.
I wrote a recipe[1] which generates a special 'Makefile'[2]. It can be
used for developing an application in a way like
| $ make -C ~/src/my-app -f
: cannot stat `.../libvpx-0.9.0/vpx-vp8-nopost-nodocs*0.9.0/*': No such
file or directory
| ERROR: Function do_install failed NOTE: package libvpx-0.9.0-r2: task
do_install: Failed
This patch makes the build more deterministic.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
Cc: Koen
-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
Cc: David-John Willis john.wil...@distant-earth.com
Cc: Martin Jansa martin.ja...@gmail.com
---
recipes/xorg-font/encodings_1.0.3.bb |4 ++--
recipes/xorg-font/font-util_1.1.1.bb |4 ++--
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git
Martin Jansa martin.ja...@gmail.com writes:
BTW: 5.4 is 6years old, 5.7 ~2years old (latest)
Is someone already working on newer ncurses recipes?
https://www.cvg.de/people/ensc/ncurses_5.7.bb
Problem with ncurses is, that it needs a patched autoconf 2.13 to do
the full autoconf step. Recipe
Henning Heinold hein...@inf.fu-berlin.de writes:
the libattr from xfsprogs clashes with the attr recipe from udev and I
can find any recipe which uses libattr, besides libcap2 which could
use attr from udev too. So my proposal is to remove the libattr recipe
and switch libcap2 to use attr
Khem Raj raj.k...@gmail.com writes:
On (15/06/10 15:03), Enrico Scholz wrote:
The st_mtime attribute (which is a float) is compared against a value
from the timestamp database, which was stored as an integer there.
Is it stored as integer or long int? Did you try it on x86_64 box.
As 'int
.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
classes/binconfig.bbclass |5 -
1 files changed, 4 insertions(+), 1 deletions(-)
diff --git a/classes/binconfig.bbclass b/classes/binconfig.bbclass
index b3b2236..c1116fc 100644
--- a/classes/binconfig.bbclass
+++ b/classes
Graham Gower graham.go...@gmail.com writes:
None of the distros in OE
At least the openmoko SHR distribution fetches ipkg from network and
requires curl support hence.
appear to want/need gpg or curl support in opkg; so have opkg.inc
remove them by default.
I do not think that supporting
Koen Kooi k.k...@student.utwente.nl writes:
A better way might be to conditionalize your '--disable-*' options
resp. make them depend on some DISTRO_FEATURE.
Introducing USE flags would be a step backwards, so NAK on that.
how would you reenable gpg and curl for https support then?
Enrico
Jan Paesmans jan.paesm...@gmail.com writes:
cp: cannot create regular file `/ld.so.conf': Permission denied
bb.decodeurl (bitbake 1.10) seems to be buggy:
| import bb
| bb.decodeurl(file://ld.so.conf)
| ('file', '', 'ld.so.conf', '', '', {})
| bb.decodeurl(file://etc/ld.so.conf)
| ('file',
Khem Raj raj.k...@gmail.com writes:
right. OE needed a fix too
http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=5ac52371b09a25518b60e23e80c8c812cdea6e36
which means that bitbake fix should also go
into 1.8 an 1.10 to keep building oe .dev with those versions of bitbake.
I still
Versions before 0.3.15 are more than 2 years old and not referenced
anywhere.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
recipes/liboil/liboil-0.3.11/ppc-detect-fpu.patch | 11 ---
recipes/liboil/liboil_0.3.10.bb | 15 ---
recipes
a negative DEFAULT_PREFERENCE
for now.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
recipes/ncurses/ncurses-5.7/tic-hang.patch | 25
recipes/ncurses/ncurses_5.7.bb | 187
2 files changed, 212 insertions(+), 0 deletions(-)
create
This makes it possible to specify whether the equivalent of 'rm -rf' or
only this of 'rm -f' is wanted. Due to backward compatibility it
defaults to the recursive variant.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
lib/oe/path.py |4 ++--
1 files changed, 2 insertions
Hi,
http://wiki.openembedded.org/index.php/Legacy_staging states that
NATIVE_INSTALL_WORKS must be set when there is a non trivial
do_install() function and BBCLASSEXTEND is used.
But
| git grep NATIVE_INSTALL_WORKS conf/ classes/ lib/
shows only one place where this variable is evaluated:
Eric Bénard e...@eukrea.com writes:
Is there a reason we're perpetuating lzo1 for newer stuff rather than
switching to lzo2?
lzo1 seems necessary :
compr_lzo.c:29:19: fatal error: lzo1x.h: No such file or directory
mtd-utils build fine here after removing *all* the lzo1 related patches
and
From: Enrico Scholz enrico.sch...@informatik.tu-chemnitz.de
When building as a dedicated user, compilation fails else in a way like
| + cd /e2fsprogs-1.41.12-r27/e2fsprogs-1.41.12
| + do_compile
| + find ./ -print
| + xargs chmod u=rwX
| chmod: changing permissions of `./patches/llseek
From: Enrico Scholz enrico.sch...@informatik.tu-chemnitz.de
@sharedlibdir@ is not substituted so that linkerflags in zlib.pc expand to
... -L@sharedlibdir@ which breaks e.g. glib builds.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
recipes/zlib/zlib-1.2.5/0001
From: Enrico Scholz enrico.sch...@informatik.tu-chemnitz.de
An umask of 022 can cause 'chmod' to fail in a way like
| $ chmod -w X
| chmod: X: new permissions are r--rw-r--, not r--r--r--
It is better to specify 'a-w' because chmod is not affected by the
umask then and changes all permission
Khem Raj raj.k...@gmail.com writes:
@sharedlibdir@ is not substituted so that linkerflags in zlib.pc expand to
... -L@sharedlibdir@ which breaks e.g. glib builds.
I think zlib has --sharedlibdir configure option since 1.2.4.4+
so you could try to use it instead of patching
afais, the
Khem Raj raj.k...@gmail.com writes:
alternatively you can exclude patches directory from feeding into
chmod
I do not think that excluding a single, implementation specific directory
is a good idea. When this or a future package contains symlinks to outside
the buildroot, problem would be
Tom Rini tom_r...@mentor.com writes:
- for PERLSCRIPT in `grep -rIl ${bindir}/perl ${D}${bindir}`; do
- sed -i -e 's|^#!${bindir}/perl|#!/usr/bin/env perl|' $PERLSCRIPT
+ for PERLSCRIPT in `grep -rIEl '#!.*/perl' ${D}${bindir}`; do
+ sed -i -e
'.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
conf/machine/include/tune-cortexm3.inc |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/conf/machine/include/tune-cortexm3.inc
b/conf/machine/include/tune-cortexm3.inc
index 6da9aee..a66a33c 100644
--- a/conf
Fix for http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43999 and is
required to build gcc for Thumb2 only architectures (e.g. Cortex-M3).
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
recipes/gcc/gcc-4.5.inc|3 ++-
recipes/gcc/gcc-4.5/thumb2-build.patch | 11
Fix for http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43999 and is
required to build gcc for Thumb2 only architectures (e.g. Cortex-M3).
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
recipes/gcc/gcc-4.5.inc|3 ++-
recipes/gcc/gcc-4.5/thumb2-build.patch | 16
Khem Raj raj.k...@gmail.com writes:
The '-mcpu=cortex-m3' is shorter than '-march=armv7-m -mtune=cortex-m3'
and enables workarounds like '-mfix-cortex-m3-ldrd'.
Does -mcpu=cortex-m3 also cause gcc to tune for cortex-m3 ?
Yes; logic in gcc/config/arm/arm.c is:
|if (i == ARM_OPT_SET_CPU
Khem Raj raj.k...@gmail.com writes:
Can you check what does gcc emits in an assembly file when using
-mcpu=cortex-m3 in your case. If it emits sdiv instruction or not
int foo (long a, long b)
{
return a/b;
}
$ arm-linux-uclibceabi-gcc -mthumb -mcpu=cortex-m3 -mfpu=vfp
to store the
fractional part too, this will complicate things more than we would
gain by this change.
This patch was applied as 49e11b32906fc1becd9218bb3125f6fc8fa03fa4 to
old ipkg-utils already.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
recipes/opkg-utils/opkg-utils/mtime
.
yes; I submitted a similar patch [1] because it is required by ncurses.
Acked-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
Enrico
Footnotes:
[1] http://patchwork.openembedded.org/patch/2248
___
Openembedded-devel mailing list
Openembedded
'-mthumb'; even when invoked with '-E':
| $ arm-linux-uclibceabi-gcc -E - /dev/null
| # 1 stdin
| stdin:1:0: error: target CPU does not support ARM mode
| $
| $ arm-linux-uclibceabi-gcc -mthumb -E - /dev/null
| # 1 stdin
| # 1 built-in
| # 1 command-line
| # 1 stdin
| $
Signed-off-by: Enrico
'-mthumb'; even when invoked with '-E':
| $ arm-linux-uclibceabi-gcc -E - /dev/null
| # 1 stdin
| stdin:1:0: error: target CPU does not support ARM mode
| $
| $ arm-linux-uclibceabi-gcc -mthumb -E - /dev/null
| # 1 stdin
| # 1 built-in
| # 1 command-line
| # 1 stdin
| $
Signed-off-by: Enrico
From: Enrico Scholz enrico.sch...@informatik.tu-chemnitz.de
openssl tries to detect multlib environments and assigns libdir=lib64
e.g. for x86_64 targets. This breaks OE assumptions about the location
of pkgconfig files and causes build failures.
As OE is not multilib aware, use always a static
From: Enrico Scholz enrico.sch...@informatik.tu-chemnitz.de
This fixes builds with '-Wl,-as-needed'. Patch changes the previous
'ldflags.patch' files instead of adding a new one.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
Cc: Jeremy Lainé jeremy.la...@m4x.org
---
.../tcp
it on the next run. Because collisions of the sha1
hash are very unlikely, the git revision is used directly as the key.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
classes/gitpkgv.bbclass | 29 +++--
1 files changed, 23 insertions(+), 6 deletions
dietlibc was splitted into target and cross recipes. These recipes
build now the latest stable version, from upstream CVS or from a github
repository with lot of fixes.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
recipes/dietlibc/dietlibc-0.32.inc | 24
After recent dietlibc refactoring, the functionality of the old
'dietlibc' recipe is now provided by 'dietlibc-cross'.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
classes/dietlibc.bbclass |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/classes
Michael Smith msm...@cbnco.com writes:
-perl ./Configure ${EXTRA_OECONF} shared --prefix=$useprefix
--openssldir=${libdir}/ssl $target
+perl ./Configure ${EXTRA_OECONF} shared --prefix=$useprefix
--libdir=lib --openssldir=${libdir}/ssl $target
Could you change lib to ${libdir}?
From: Enrico Scholz enrico.sch...@informatik.tu-chemnitz.de
openssl tries to detect multlib environments and assigns libdir=lib64
e.g. for x86_64 targets. This breaks OE assumptions about the location
of pkgconfig files and causes build failures.
Patch gives the directory name to openssl which
Paul Menzel paulepan...@users.sourceforge.net writes:
From: Enrico Scholz enrico.sch...@informatik.tu-chemnitz.de
it seems you need to update your Git configuration to your common
email address.
Both addresses are in use; some patches are created at home when I work
on a personal project
Koen Kooi k...@dominion.thruhere.net writes:
+perl ./Configure ${EXTRA_OECONF} shared --prefix=$useprefix
--libdir=`basename ${libdir}`
Why not plain ${base_libdir}?
openssl's '--libdir' option expects a relative name. E.g. there is done
| $libdir=lib$multilib if $libdir eq ;
| ...
|
Paul Menzel paulepan...@users.sourceforge.net writes:
For documentation purpose it would be great if you could add such
acknowledgments to the commit message.
ok; I will add it the next time.
But I meant upstream of TCP Wrapper.
tcp-wrappers have an upstream? afais, the tarball in the
Jaap de Jong jaap.dej...@nedap.com writes:
TARGET_ARCH = x86_64
did this ever work? OE uses hardcoded 'lib' (instead of 'lib64') at
various places. At least gcc has
| #define GLIBC_DYNAMIC_LINKER64 /lib64/ld-linux-x86-64.so.2
which causes 64 bit binaries to look for the interpreter at
Phil Blundell ph...@gnu.org writes:
x86_64 seems to be what's currently used as the architecture name. If
you want a value for ${MACHINE} then I guess something like
x86_64-generic would do. I don't think there's any prohibition
against underscores in machine names.
OVERRIDES don't work
Michael Smith msm...@cbnco.com writes:
did this ever work? OE uses hardcoded 'lib' (instead of 'lib64') at
various places.
This is my MACHINE config:
...
# Some packages install files into /lib/packagename, not /lib64/packagename.
FILES_${PN} += ${prefix}/lib/${PN}/*
FILES_${PN}-dbg +=
Martin Jansa martin.ja...@gmail.com writes:
Is it possible to implement some support for old functionality like
bitbake -c clean -b path.to.recipe.with.BBCLASSEXTEND.bb
...
It's really hard to implement something like prefixing -b path with
virtual:native:
Bitbake must be patched for this;
:
| oe_runconf --with-build-cflags=${BUILD_CFLAGS}
This patch might create incompatibilities for recipes which workaround the
current limitations.
As a sideeffect, 'set -x' debug output is now much nicer because
superflous whitespaces in the configure call are removed.
Signed-off-by: Enrico Scholz
on absolutes paths
which are readded by this patch.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
classes/kernel.bbclass | 11 +++
1 files changed, 11 insertions(+), 0 deletions(-)
diff --git a/classes/kernel.bbclass b/classes/kernel.bbclass
index 69ab422..29a7e64 100644
directory instead of the single files.
Signed-off-by: Enrico Scholz enrico.sch...@sigma-chemnitz.de
---
classes/packaged-staging.bbclass | 21 -
1 files changed, 16 insertions(+), 5 deletions(-)
diff --git a/classes/packaged-staging.bbclass b/classes/packaged-staging.bbclass
Paul Menzel paulepan...@users.sourceforge.net writes:
This patch allows to pass quoted strings in EXTRA_OECONF. E.g. with
this patch, it is possible to do
| EXTRA_OECONF = --with-build-cflags='${BUILD_CFLAGS}'
where 'BUILD_CFLAGS' contains multiple, whitespace separated arguments.
...
Tom Rini tom_r...@mentor.com writes:
On Tue, 2010-03-16 at 13:23 +0100, Enrico Scholz wrote:
Doing a '-c clean' operation on a staged package with very much files
(e.g. glibc) took several minutes because
* every removed file was reported
* an 'rm' instance was spawned for every file
Martin Jansa martin.ja...@gmail.com writes:
Is it possible to implement some support for old functionality like
bitbake -c clean -b path.to.recipe.with.BBCLASSEXTEND.bb
...
It's really hard to implement something like prefixing -b path with
virtual:native:
Bitbake must be patched for
From: Enrico Scholz enrico.sch...@informatik.tu-chemnitz.de
Old .inc file listed every single header which was to be staged. This
causes problems when upgrading to recent 0.5.1 version or when using
older versions because headers might not exist there resp. new ones
need to be installed
Using an 'rm -rf $dir' ordering is more portable.
Signed-off-by: Enrico Scholz enrico.sch...@informatik.tu-chemnitz.de
---
classes/rm_work.bbclass |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/classes/rm_work.bbclass b/classes/rm_work.bbclass
index a53d12b..0857f49
Eric Bénard e...@eukrea.com writes:
-'narrowc --with-ticlib' \
-'widec --enable-widec --without-progs'; do
+'narrowc --with-ticlib' \
+'widec--with-ticlib --enable-widec --without-progs'; do
can you place this in a separate patch please?
+ cd
Eric Bénard e...@eukrea.com writes:
+RDEPENDS_alsa-utils-alsamixer = ncurses-terminfo
I think this should be solved in ncurses because it affects other
packages too; e.g. moving its /etc/terminfo files from 'ncurses' into
an own package and rrecommending this in the libtinfo package.
Enrico
Martin Jansa martin.ja...@gmail.com writes:
+RDEPENDS_alsa-utils-alsamixer = ncurses-terminfo
I think this should be solved in ncurses because it affects other
packages too; e.g. moving its /etc/terminfo files from 'ncurses' into
an own package and rrecommending this in the libtinfo
Chase Maupin chasemaupi...@gmail.com writes:
* Added patch to allow using older versions of automake such
as 1.10.3 when building jpeg version 8b.
* Only use the AM_SILENT_RULES if it is available.
Signed-off-by: Chase Maupin chase.mau...@ti.com
Acked-by: Enrico Scholz enrico.sch...@sigma
Eric Bénard e...@eukrea.com writes:
-'narrowc --with-ticlib' \
-'widec --enable-widec --without-progs'; do
+'narrowc --with-ticlib' \
+'widec--with-ticlib --enable-widec --without-progs'; do
can you place this in a separate patch please?
I'm not
Eric Bénard e...@eukrea.com writes:
5.7 is also broken for sdk build,
I think that I fixed widec builds for native targets in git. 'sdk' and
'native' works on Fedora 13 now; please check other distributions.
Enrico
___
Openembedded-devel mailing
Jaap de Jong jaap.dej...@nedap.com writes:
$ bitbake cmake-native
| [ 87%] Building CXX object
Source/CMakeFiles/CPackLib.dir/CPack/cmCPackNSISGenerator.o
| /usr/bin/ld: cannot find -ltinfo
| collect2: ld returned 1 exit status
| make[2]: *** [bin/ccmake] Error 1
| make[1]: ***
Frans Meulenbroeks fransmeulenbro...@gmail.com writes:
Eric peeked into the termcap problem. I've tested his patch and it
works. He will probably push a patch later today.
The termcap problem from asterisk is not related. It cannot find
libtermcap. This one does not exist any more with
Tom Rini tom_r...@mentor.com writes:
| /usr/bin/ld: cannot find -ltinfo
no, 'DEPENDS = ncurses-native'; BUT:
* it will cause problems on (host)distributions which do not have libtinfo
(e.g. RHEL5) and have ncurses-native in their ASSUME_PROVIDED
Isn't the whole point of ASSUME_PROVIDED
Tom Rini tom_r...@mentor.com writes:
I added 9e58422868a03eeafa72781477174850d1c5abe7 some hours ago so
that '-ltermcap' should work again. Patch does not break my builds
but I do not know whether it unbreaks the problematic ones ;)
Things are too different across the distributions (some
Hi,
in the last days I committed several changes to the ncurses receipt
which should fix:
* broken widec builds for native/sdk targets
* missing -ltermcap
* implicit -ltinfo on -lncurses (-- this might require a
| bitbake ncurses -c clean
to take an effect because staging does not
Martin Jansa martin.ja...@gmail.com writes:
+# actually RDEPENDS, but ncurses-terminfo is only in ncurses-5.7, so
RRECOMMENDS for now
+RRECOMMENDS_${PN} = ncurses-terminfo
...
Is there something special in ncurses-terminfo 5.7 that nano needs?
There wasn't any common /etc/terminfo in base
Holger Freyther holger...@freyther.de writes:
when building for XScale one could pass -iwmmxt to the configure
script.
I would avoid -march=iwmmxt because
a) there seems to be kernel bug which does not restore IWMMXT state on
ELF loading[1]
b) gcc 4.4.4 is bugged for this architecture
Enrico Scholz enrico.sch...@sigma-chemnitz.de writes:
Probably not on ARMv5 ;) There 5/3 == 3 holds and ld.so crashes due to
related errors.
sorry; my fault. I applied a broken patch from [1].
Enrico
Footnotes:
[1] http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43999
Paul Menzel paulepan...@users.sourceforge.net writes:
- install -m 0755 src/cacaoh/.libs/cacaoh ${D}${bindir}/cacaoh-${PV}
+ install -m 0755 src/cacaoh/cacaoh ${D}${bindir}/cacaoh-${PV}
Location of the binary depends on libtool version and other factors so
that both paths are possible.
|
Frans Meulenbroeks fransmeulenbro...@gmail.com writes:
- os.system('cp -pPR %s/* %s/' % (dest, dvar))
+ os.system('cp -pPR %s/* %s/.??* %s/' % (dest, dest, dvar))
why not os.system('cp -pPR %s/. %s/' % (dest, dvar))?
Enrico
___
1 - 100 of 113 matches
Mail list logo