The current free space doesn't leave much room for using the image,
increase it. Patch from Saul Wold with tweaks.
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
Per bpo#109 dependency failure during rootfs generation did not cause an error
despite that it would cause errors later on in the process. This patch updates
the key phrases searched for in the log checking function.
Signed-off-by: Joshua Lock <josh@linux.intel.com>
this is one existing for stale reason when package's own Makefile doesn't
handle .pc correctly. After previous commits to remove unnecessary inheritage
on it (sqlite3, libpng), the only remaining is libgcrypt. Libgcrypt itself
doesn't ship a .pc and poky generates one. In this case, simply move the
manual installation to its own do_install.
Then it's safe to remove this class file
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
Since scripts is now in PATH thanks to the layer functionality there is
no longer any need to have this recipe full of special cases, the scripts
can just be placed there.
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
NATIVE_INSTALL_WORKS implicates an explicit do_install required or else
its binaries are not installed to sysroot, which result in error like:
sh: opkg-build not found
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
[Patches]
KEEP _ignore_arch_directory_: ignore /etc/modprobe.d/arch, not sure reason now.
keep it and to be verified later
REMOVE _module-init-tools-remove-index.patch_: in upstream
NEW _disable_man.patch_: a simpler patch to replace below removed two
REMOVE _manpagesopt_: due to above new patch
REMOVE _no_man_rebuild_: ditto
KEEP _modutils_extension_: poky has its own name convention and thus need such knowledge
in code
[Recipe]
- adjust indention
- remove some old lines like ${S} and EXTRA_OECONF
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
although the latest version is 1.4.3, some API changes happen which are not
compatible to some packages (such as libmatchbox using png_check_sig which
is abandoned in 1.4 version). I didn't check fully about other broken places,
and then be conservative to choose 1.2.44 as the target instead atm.
[Patches]
REMOVE _makefile_fix.patch_: a specific fix for libtool2.2.2. with 2.2.10, it's
not required then
[Recipe]
- add license checksum
- remove old do_install and do_package
- clean up package split
- remove inherit pkgconfig_stage, since its own Makefile handles it well
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
libidn changed to GPLv3 since 2007 (its 1.0 release), so this upgrade includes two:
0.6.14 (GPLv2):
- change SRC_URI as original doesn't work
- remove original do_configure_prepend, which doesn't make any sense when inheriting
autotools
- this version includes a local copy of AC_USE_SYSTEM_EXTENSIONS which conflicts with
latest autoconf. Actually it's not necessary and remove all references to local version
1.19 (GPLv3):
- split a new package 'idn' for command line tools, which are licensed under GPLv3.
keep other bits in libidn licensed under LGPLv2.1
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
[Patches]
REMOVE _autoreconf.patch_: no need to duplicate acinclude.m4 in every subdirectory
[Recipe]
- Keep original 3.0.3 version for GPLv2 requirement. 3.0.4 has bumped to GPLv3+
- add gperf.inc for common lines
- instead of duplicating acinclude.m4 in every subdirectory, use " -I ${S}" to
ensure acinlude.m4 successfully found by m4_include
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
[Patches]
KEEP _unlinux.patch_: remove reference to build system paths
DISABLE _libtool.patch_: this patch adds a "--tag=CC" to be compatible with libtool2.2.4.
however with latest 2.2.10 libtool, w/o this patch it still works. From
the manual, CC is the default tag actually. So disable it for now, and once
same error happens again, it'll be re-neabled and if necessary push to upstream
DISABLE _workaround.patch_: a SIGUSR1 is hooked to signal suspend event as a so-called 'workaround'.
however no exact commit is found for exact usage case. So disable it.
REMOVE _debian.patch_: in upstream
[Recipe]
Add license checksum
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
This package changed from autoconf to cmake during the last recipe
upgrade. This made the existing EXTRA_OECONF, which disabled wekbit
support among other things, non-functional, which in turn, broke the
build.
This patch disables webkit in a CMAKE way by adding EXTRA_OECMAKE to
the cmake class description, and setting it to "-DWEBKIT=no" in the
libproxy recipe and also reproduces the EXTRA_OECONF settings.
Signed-off-by: Jeff Dike <jdike@linux.intel.com>
reuse the open embedded elfutils_0.143.bb recipe to fix
the cross compiling issue.
Also pick patches from debian elfutils-0.148-1 pkg
Signed-off-by: Yu Ke <ke.yu@intel.com>
Update the default revision for opkg-utils in poky-default-revisions.inc
Remove "inherit autotools" since opkg-utils is not autotools based
Fix recipe metadata
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Use stable release as default recipe.
Remove logfix.patch since the logic is already in latest version package
Also fix the metadata
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
gpgme 1.3.0 requires gpgconf&g13 from gnupg, otherwise got configure failure.
Adding gnupg need a lot of efforts and still has issue:g13 only exist in
development branch. So add a patch to disable the check error, just like
non-cross-compile environment.
Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
font-alias: change the RECIPE_STATUS to green as the license issue is clarified.
liberation-fonts: add "# OK" to indicate we don't update it from 1.04 to 1.05.2.20091019 for now.
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
Change
RDEPENDS_task-poky-tools-profile_qemux86 += "valgrind"
to
RDEPENDS_append_task-poky-tools-profile_qemux86 = "valgrind", or else
RDEPENDS_task-poky-tools-profile would be overriden and only holds valgrind,
as a result, the 5 packages oprofile/oprofileui-server/powertop/... aren't
built into the target image at all.
Thanks Kevin a lot for tracking the issue down.
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
Original patch status:
ldflags.patch:
paths.patch:
libtool_tweak.patch:
Dropped as the issues got fixed in 8b
debian.patch:
This patch comes from debian distro. Original version can't apply, so
replace it by latest version
Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
Also backport a patch from upstream(f001953f7d60) to compile imapx provider
optionally, which is the default behavior of current version. Else we get build
error when compiling imapx.
Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
Uprev pseudo to the latest head version to resolve a number of issues
on older hosts such as RHEL 5. In addition sqlite was changed to link
statically into pseudo to avoid a potential LD_LIBRARY_PATH issue.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
You need to first set up the build directory by sourcing the poky build script,
after that builds can be run in that directory so long as bitbake is in $PATH
removing the need to source the init script for each build.
i.e:
$ . poky-init-build-env ~/my-build
$ bitbake some-image
<<later, in a different shell>>
$ cd ~/my-build
$ export PATH=/path/to/bitbake/bin:$PATH
$ bitbake an-image
This patch also removes use of OEROOT in recipes, etc.
Signed-off-by: Joshua Lock <josh@linux.intel.com>
Having PSTAGE_DIR be a top level directory by default doesn't make sense, move
it to be a child of the build directory so that it lives with all other built
output. Also move DL_DIR to a child of the build directory for the use case of
an unwriteable $OEROOT.
Signed-off-by: Joshua Lock <josh@linux.intel.com>
This base recipe implements support for building a git based
Wind River Linux kernel repository.
There are several restrictions in this first import:
- no patching (needs guilt)
- missing Wind River features (config audit, etc)
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
These tools are largely unused at the moment, but represent
a baseline set of functionality for working with a Wind River
Linux kernel.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Put PREFERRED_PROVIDER_linux-libc-headers = "linux-libc-headers-wrs"
in your configuration, and this package will be used to generate a
set of kernel headers from the -standard branch of the Wind River kernel.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Now perl version is in two digits like 12 in 5.12.1. Fix the old
function to work with multi digit version strings.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
new upgrades:
genext2fs
pcmciautils
pkgconfig
shared-mime-info
add missing:
util-linux
libusb1
misc:
latest version typo in libusb
add # OK flag to openssl
clarify tinylogin
Signed-off-by: Qing He <qing.he@intel.com>
from 2.6.9
2.6.9 is also preserved as the last known GPL2 version
changes:
- 3.0.7 has some new change to the autotools scripts, it uses
and configure.sh and checks its existence in Makefile, adapt
this by adding do_configure_{prepend,append}
- aclocal.m4 now also contains custom m4 macros, add acinclude.m4
in the files/
- add popt as build dependency. Besides, rsync can also link to
libacl and libattr. Possibly need to add them in the future, for
now, document the dependency
Signed-off-by: Qing He <qing.he@intel.com>
from 014
changes:
- put LEX=flex into EXTRA_OEMAKE, from openembedded project
- add RDEPENDS only to ${PN}, from openembedded project
- remove modalias_update patch, it's in upstream
- rebase the other two patches
- starting from 017 (or maybe 016), two helper binaries are put
into /usr/lib/udev instead of /sbin, adapt this change by adding
the new path into package
Signed-off-by: Qing He <qing.he@intel.com>
from 0.23
changes:
- remove sysrootfix patch, it's now in upstream
- use external popt instead of the included one
Signed-off-by: Qing He <qing.he@intel.com>
from version 1.3
changes:
- autotools adaption
- remove the patches since their base 1.3 diverges considerably
from the latest version
Signed-off-by: Qing He <qing.he@intel.com>
Removed local patches since the changes have accpeted in latest version
Rename wpa-supplicant-0.5.inc to wpa-supplicant-0.7.inc
Modify ${S} since the new code structure is changed
Also fixed recipe metadata
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Remove the patch since build tool has changed from autotool to cmake
Export poky variable to system environment since cmake needs it
Also fix the metadata
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
libxsettings-client: the latest version should be "0.10" rather than "0.1";
font-util: the license has been updated;
xorg-minimal-fonts: the license has been updated;
randrproto: fix the latest version: it should be "1.3.1" rather than "0.11";
libxfont: the package has been upgraded from "1.4.1" to "1.4.2" recently;
libxcalibrate: fix the git version's format.
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
There is no explicit license info in the local directory.
I used "fontforge" to open the fontdata file (.pcf.gz file) in the
xorg-minimal-fonts/misc/ and found fontforge's menu
"Element->Other Info->BDF info" says the "COPYRIGHT" is
"Public domain font. Share and enjoy. Actually
"cat *.pcf.gz | gunzip | strings | grep 'Public domain'" also shows
the same string.
So I think the license of xorg-minimal-fonts should be Public Domain.
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
Pseudo defaults to storing state files in ${prefix}/var/pseudo, we want them in
$(localstatedir) so a quick hack (path-munge.patch) makes pseudo use a data
directory specified with --data, and defaults to pseudo's way if it's not set.
Touching LD_LIBRARY_PATH can confuse the system into running Python against a
staging python library. When these two are sufficiently diverse in version
significant breakage can occur.
HOMEPAGE and LIC_FILES_CHKSUM are handy metadata variables to have defined.
Signed-off-by: Joshua Lock <josh@linux.intel.com>
If a program creates an empty directory during its build it's probably for a
good reason, so if it exists it seems sensible to stage it.
Some programs require localstatedir to operate so we should be sure to stage
it.
Signed-off-by: Joshua Lock <josh@linux.intel.com>
As per discussion with RP remove dolt patches.
Merge Gary Thomas's patch into trailigslash.patch
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
Now Warning includes the package name like zlib is printed bellow:
WARNING: zlib: Recipe file does not have license file information
(LIC_FILES_CHKSUM)
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
The sulogin compile checked for /usr/lib*/libcrypt.a to decide whether
to add -lcrypt to the final compile. However, the recipe puts LCRYPT
in the environment, so crypt-lib.patch makes the Makefile check for
this before looking at the host /usr/lib.
Signed-off-by: Jeff Dike <jdike@linux.intel.com>
Automation test is disabled by default. User need set TESTCLASS
to qemu in conf/local.conf and run bitbake command "bitbake
poky-image-xxx" or "bitbake poky-image-xxx -c qemuimagetest" to
trigger it. Currently only the sanity test with two testcases are
added.
To run the test, user need prepare a testing environment:
1) "expect" should be installed on system
2) NOPASSWD should be set for user to run bitbake
Signed-off-by Jiajun Xu <jiajun.xu@intel.com>
In order to force 0.12.4 to allow a powerpc system boot, we need
a patch to qemu, and a new powerpc (not ppc) ROM.
This configuration supports non-graphical booting only, since the
ROM does not have support to pass a graphics device via the
device tree yet.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Current binutils generates problematic relocation table for executable
compiled with '-fPIE', such as dbus-daemon. That way ld.so.1 falls into
segmentation fault when accessing bad reloc entry. This pulls back one
fix from binutils CVS repo (1.267) to add PIE support back.
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
Only checking the license at packaging time means we don't check native builds,
so move the check to after do_configure
Signed-off-by: Joshua Lock <josh@linux.intel.com>
Override PACKAGES_DYNAMIC in gdk-pixbuf-csource-native_2.12.7.bb since
it requires gtk+_${PV}.bb where PACKAGES_DYNAMIC has already been
defined
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Rename all the sub-plugin to eliminate the inclusion by all plugins set
This can fix the multple providers warning
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Fix a missing config.h dependency when building sparseset.o to prevent
"conflicting types for 'strsignal'" error, which can intermittently crop
up for high values of make -j.
Thanks to Richard Purdie and Jeff Dike for their help in tracking this
down.
Signed-off-by: Scott Garman <scott.a.garman@intel.com>
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
Some packages are not feasible for automatic version check. This commit
adds a weak list of version information for those packages before they
are carefully auditted.
Also there're 14 packages out of tracking so far. Add simple information
for them too.
This is the last one to be assorted, so use weak assignment "?="
Signed-off-by Kevin Tian <kevin.tian@intel.com>
along with recipe upgrade, previous identified issues are solved, and thus
it's time to update status now:
expat
dbus
dbus-glib
busybox
prism-firmware
base-files
base-passwd
module-init-tools
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
[dbus-glib-Ptches]
REMOVE _run-with-tmp-session-bus.patch_: test is not built by default
DISABLE _fix_asneeded.patch_: not sure why we need change expat link order. Disable it
but keep for a while
KEEP _no-examples.patch_: don't build examples
[dbus-glib-recipe]
- remove native and old recipes
- remove tweak on dbus-glib-bindings.h, since it will be generated correctly after passing
correct config options linking to native dbus-binding-tool and dbus-bus-introspect.xml
- remove preferred versions in poky-fixed-revisions.inc with similar reason as dbus
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
[Patches]
REMOVE _cross.patch_: fixed in upstream
KEEP _tmpdir.patch_: disable tmpdir test in build system
UPDATE _fix-install-daemon.patch_: use automake for binary install instead of manual
tweaks. extend it to cover dbus-daemon-launch-helper
[Recipe]
- remove native recipes and old versions
- install to STAGING_DATADIR_NATIVE directly
- remove manual tweak on dbus-daemon-launch-helper install
- remove stale config options
- adjust indention
- enable dbus in runlevel 3 since networkmanager has requirement at that level
- remove preferred versions in poky-fixed-revisions.inc. Based on discussion with Josh,
the preferred version is there just because dbus is not carefully tested before
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
gtk+ previously has 2 versions: 2.14.2(preferred), and 2.18.5(latest). This
upgrade is from 2.18.5 to 2.20.1, so some patch(like entry-cairo.patch) in
2.14.2 is abandoned.
Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
Richard Purdie
Joshua Lock
Kevin Tian
Dexuan Cui
Yu Ke
Dongxiao Xu
Zhai Edwin
Jeff Dike
Mark Hatle
Bruce Ashfield
Nitin A Kamble
Qing He
Jiajun Xu
Scott Garman