package_deb.bbclass: make the PACKAGES check earlier so if there is nothing to package we don't read potentially stale pkgdata from staging.
This is basically the commit 27b9bc946eeb3faa7b10d92689ab527af42727bd that got
lost when merging poky stuff into the repo. The behavior is much like
what we already have with package_ipk.
Signed-off-by: Ricardo Salveti de Araujo <ricardo.salveti@openbossa.org>
Roman Khimov [Wed, 15 Jul 2009 02:52:41 +0000 (02:52 +0000)]
icecc: introduce ICECC_FALLBACK_PARALLEL
icecc-blacklisted packages currently built not only without icecc environment
set but also without any parallel options enabled. The original intention was
not to kill the build machine, I guess, since having a good icecc cluster it's
easy to set '-j16' or '-j32' or whatever else which could do so. However as
the toolchain packages are blacklisted for icecc they're becoming a
bottleneck in the build process, while they can be built with local-machine
parallelism.
So, in order to solve it, introducing ICECC_FALLBACK_PARALLEL option. It's
simply used as PARALLEL_MAKE for icecc-blacklisted packages.
Roman Khimov [Wed, 15 Jul 2009 11:52:01 +0000 (11:52 +0000)]
icecc: remove bb.errors on un-built compilers
Building things with icecc and BB_NUMBER_THREADS more then 1 you could see
a lot of messages:
ERROR: no cross compiler built yet?
before gcc-cross would be properly built. There is technically nothing wrong,
cross compiler is not there yet and it's not needed when bitbake does
fetch-unpack-patch and other things to prepare package build. But it makes
people uncomfortable.
Without the gtkmathview dependency the abiword mathview plugin is not
built,which prevent the edition of the math formulas(mathml or latex
formulation)
I also had issues displaying the formulas on my eeepc 701
(machine support will be commited soon) without the plugin:
*The 3 top formulas of my document had a black rectangle on a part of
them
*the part after the black rectangle was truncated and mixed
Installing the plugin fixed both issues,but I noticied an issue
with the latex notation where a \cdot was transformed in a C with
a vertical bar in it(like the $)
See https://bugzilla.gnome.org/show_bug.cgi?id=444974 for more details
I encountred this bug when trying to run an application, that made use
of the 2 screens of the bug device(see bug.conf in machines dir)
The patch is already in the trunk of pygtk,and so is already upstream
I took the default config and changed the following things:
hda1->sda1: eee710 uses the new pata/sata subsystem driver in
the linux kenrel,and the devices nodes changed from hdxx
(in use in the old ata subsbsystem) to sdxx (in use in the new
unified pata/sata one)
ro was temporarly remplaced by rw because of the problem where the boot
process hung.
Udev could not write his cache on the disk, should have been fixed
by commit 2b835db69e6b838fa8628dd85b11b4f930d5f5f9 but it was not
(same issue than before...maybe this commit fix something else then)
The eeepc 701 has a synaptics touchpad,so we use the synaptics
driver which gives us more features and is more convenient
to use,specially with the TapButton1 option,which permit
to emulate the fact of holding the mouse button while moving
Even if evdev was in the kernel,evdev didn't seem to work for:
*internal keyboard
*external keyboard
*external mouse
So it uses the mouse and kbd driver
Using standard vanilla linux kernel because everything is supported
in it
Also using the 2.6.32 version because 2.6.33 is not out yet and I
only made a defconfig for 2.6.32
added MACHINE_DRI_MODULES = "i915" instead of "intel"
because intel didn't work:
| make[5]: Entering directory `.../Mesa-7.6/src/mesa/drivers/dri/intel'
| make[5]: *** No targets specified and no makefile found. Stop.
Copied the glibc part from ion.conf(I didn't test it yet)
I also added the xf86-input-synaptics at the last moment and I didn't
have the time to test that modification but I tested
xf86-input-synaptics and it works fine
It also uses xf86-input-mouse and xf86-input-keyboard as xorg.conf will use them
(evdev didn't work for me even if it was compiled in the kernel,there was
xf86-input-evdev installed and there was an entry in xorg.conf)
Here how it was configured:
*All internal hardware support was built-in
*All external hardware which could be possibly attached was compiled as module
*All hardware which can't be attached such as other pci cards than the builtin
ones are removed(note that the eeepc 701 doesn't have pc-card)
*The net filtering parts (needed for iptables) were compiled as modules when
possible
Hardware details:
*The sd-card controler uses the mass storage driver
*The synaptics driver was built as evdev and would need the synaptics xorg
driver for working,Also note that it can also function as a basic mouse
but it wouldn't benefit from the synaptics features
*The webcam is uvc
*the network card needs the atl2 driver to work
*the wifi driver is ath5k
*there is an eeepc driver for hotkeys in the kernel named eeepc-laptop
*the graphic card is an intel i915 card
Chris Larson [Wed, 6 Jan 2010 23:32:40 +0000 (16:32 -0700)]
iptables: unbork the 1.3.8 build, merge iptables-utils into iptables.
In recent iptables versions, the -save and -restore files are just symlinks
pointing to the main, merged binary, so the separate package is a bit
unnecessary.
Guo Hongruan [Thu, 21 Jan 2010 22:55:56 +0000 (22:55 +0000)]
Fix a bug of COMPATIBLE_HOST defination in linux-sun4cdm_*.bb
* Correct the defination of COMPATIBLE_HOST in linux-sun4cdm_2.4.26.bb and linux-sun4cdm_2.6.8.1.bb, the previous defination causes linux-sun4cdm never been selected.
Signed-off-by: Marcin Juszkiewicz <marcin@juszkiewicz.com.pl>
Marc Olzheim [Fri, 22 Jan 2010 08:22:35 +0000 (00:22 -0800)]
iproute2: Remove unneeded bash dependency
* Even though the recipes did not list it in REDEPENDS, two scripts in
this package depended on bash. Since this was unneeded, I removed that
dependency. (Upstream notified)
Bernhard Kaindl [Sun, 22 Nov 2009 01:22:28 +0000 (01:22 +0000)]
htop_0.7.bb: Fix htop do_configure: Ensure that ncurses is staged.
* htop's do_configure needs ncurses, but htop did only have an RDEPENDS
on ncurses which only triggers a build of of ncurses after or in parallel
to htop's do_configure, which is too late. Fix this:
Since htop is dynamically linked against ncurces, OE detects the REDEPENDS
to ncurses automatically and we can simply change the RDEPENDS to DEPENDS.
Signed-off-by: Bernhard Kaindl <bernhard.kaindl@gmx.net> Acked-by: Leon Woestenberg <leon@sidebranch.com> Signed-off-by: Khem Raj <raj.khem@gmail.com>
Bernhard Kaindl [Sun, 22 Nov 2009 02:19:14 +0000 (02:19 +0000)]
binutils_2.17.bb: Fix setscene of binutils-cross-sdk_2.17, switch binutils_2.17.bb to INC_PR
* All of the binutils-cross-sdk_*.bb recipes rely on their binutils_*.bb
recipe to define INC_PR, but binutils_2.17.bb does not define it, so
setscene of binutils-cross-sdk_2.17 fails because it does a mkdir of
work/.../binutils-cross-sdk-2.17-\$\{INC_PR\}.1/ which causes an abort
when trying to find the setscene scripts later.
Fix this by switching binutils_2.17.bb to INC_PR like silbings have done.
Signed-off-by: Bernhard Kaindl <bernhard.kaindl@gmx.net> Signed-off-by: Khem Raj <raj.khem@gmail.com>
Jan Paesmans [Mon, 21 Dec 2009 23:05:29 +0000 (23:05 +0000)]
gSOAP
Hello,
In the gSOAP package I noticed the following patch was applied to the
native build: use-just-built-binary.patch
This patches the makefiles of gSOAP so that the correct executable is used.
However the problem originates from the fact that the makefile
variable $(SOAP) is overridden in gsoap_${PV}.bb with the following
directive:
Eric Bénard [Sun, 17 Jan 2010 10:32:24 +0000 (10:32 +0000)]
barebox: Add recipe for this new bootloader
* barebox is "a bootloader that inherits the best of U-Boot and the Linux kernel"
* add recipe for v2009.12.0 (last stable)
* add recipe for git repository
Signed-off-by: Eric Benard <eric@eukrea.com> Signed-off-by: Khem Raj <raj.khem@gmail.com>
in `local.conf` made `do_compile` error out with the following message [1].
arm-angstrom-linux-gnueabi-ld -soname libcap.so.1 -x -shared -o libcap.so.1.10 cap_alloc.o cap_proc.o cap_extint.o cap_flag.o cap_text.o cap_sys.o
arm-angstrom-linux-gnueabi-ld: cap_text.o: relocation R_ARM_MOVW_ABS_NC against `a local symbol' can not be used when making a shared object; recompile with -fPIC
Khem Raj explained it the following way [2].
Am Mittwoch, den 13.01.2010, 02:27 -0800 schrieb Khem Raj:
> > I also get the same error which I did not get before with GCC 4.1.1.
>
> Older versions of binutils/ld can not detect MOVT/MOVW relocations which
> are absolute and cant be used in PIC code.
> hence even if the error was there it remained latent.
Therefore just append `-fPIC` to the CFLAGS and bump the PR variable.
The strange thing is still, that other distributions seem to have patched libcap to build with `-fPIC` long ago [3–5].
Guo Hongruan [Sun, 17 Jan 2010 13:10:18 +0000 (21:10 +0800)]
Set preferred glibc to 2.10.1 in conf/distro/micro.conf
* the preferred version of binutils of conf/distro/micro.conf is 2.19.51
* the preferred version of glibc is 2.6.1 which is defined in sane-toolchain.inc
* the ld produced by binutils-2.19.51 reports that it can not find the __begin symbol when linking with glibc-2.6.1
Stanislav Brabec [Wed, 20 Jan 2010 20:54:25 +0000 (20:54 +0000)]
xserver-xorg: Worked around server termination after chvt on some fbdev devices.
* The patch is just a quick hack that fixes chvt on some fbdev devices.
* For more read recipes/xorg-xserver/xserver-xorg/hack-fbdev-ignore-return-mode.patch
projects / openembedded.git / log