Move function rootfs_xx_log_check() from rootfs_xx.bbclass to relevant
package_xx.bbclass. (Where xx is rpm/ipk/deb).
Signed-off-by: Lianhao Lu <lianhao.lu@intel.com>
As discovered by Lianhao Lu and Kevin Tian:
"""
do_package checksum changes with switching the order In PACKAGE_CLASSES, e.g:
From
PACKAGE_CLASSES = "package_rpm package_ipk"
To
PACKAGE_CLASSES = "package_ipk package_rpm"
This is undesired since the order only matters for final rootfs generation. The actual
cause is related to how do_package is generated, which depends on when
package.bbclass is first brought in.
"""
Since we never need the EXPORT_FUNCTIONS funcationality for do_package we just drop
it which removes this problem and stablisised checksums.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
'nostamp' is normally used for some standalone tasks like clean, checkuri, etc.
It doesn't make sense to mark do_populate_lic as 'nostamp', which is depended
by do_package. No stamp file in the build dependency chain implicates that
do_package needs to be reinvoked and thus further do_package_write* stuff.
This together with another sstate bug fully confused recent master, that people
keep observing unnecessary do_package rebuilt with or without change. Below is
a short explanation for two behaviors we observed:
a) a fresh build, and then bitbake same target w/o any change. User will observe
one unnecessary rebuild and then later rebuild is fine:
[1st rebuild]
* do_populate_lic has no stamp, and thus do_package is thought not current
* do_package_setscene is then invoked with stamp file created
* then later do_package_write*** also are rebuilt
[2nd and later rebuild]
* do_populate_lic has no stamp, and thus do_package is thought not current
* do_package_setscene has stamp file and thus no need to re-execute setscene
* thus no further rebuild required
b) a fresh build, and then adjust order of PACKAGE_CLASSES. There's one bug
regarding to PACKAGE_CLASSES and do_package_setscene, that only switch the
order in PACKAGE_CLASSES generates different checksum. In this case user may
observe up to 3 rebuilds when switching package_rpm/package_ipk back and forth.
Signed-off-by: Lianhao Lu <lianhao.lu@intel.com>
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
The class strips out operators from the LICENSE field and to reduce it to a
list of license, | was missing but is quite frequently used so add it to the
list of replaced characters.
Signed-off-by: Joshua Lock <josh@linux.intel.com>
The package-split summaries were being pulled in from the main package, not
the split package metadata.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
This includes two method for build rpm repo:
1. create the metadata in rootfs_rpm
2. standalone binary for building the metadata
Not both of them are needed, generally #2 fits more for the purpose,
but #1 may have its use on rootfs creation using zypper.
Both share some problems and are subjected for future improvement:
1. the createrepo now builds metadata for the whole directory,
if there are more than one arch, it builds for all, which means
rootfs_rpm may run longer if more builds have been run.
2. createrepo builds metadata for stale rpms
Signed-off-by: Qing He <qing.he@intel.com>
The RPM platform configuration file lists all of the supported architectures,
while RPM itself doesn't use this information, other tools may use it to
determine compatibility. The first item is the default architecture for a
system, while the following items specify alternative compatible architectures.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
This is an intial commit for the license reporting system. A few notes:
The LICENSE fields needs to be standardized throughout poky. As it
stands, we throw a warning if the license file is not found (either
because it does not exist or because LICENSE_FILE_CHKSUM is munged)
in the generic license directory. This should eventually become an
error.
I've seen a few places where Apache-v2.0 is written differently and
I'm sure this will throw the above warning. This does not put the
license data on the rootfs. Also, I provide both the actual license
text and a link to the best guess of the generic_license. That guessing
is not very robust and I'm loath to get into a bunch of pattern matching
rather than standardize LICENSE.
This adds one new param to poky.conf and one new to license.bbclass:
LICENSE_DIR: the base directory we copy all the license results to (set
in license.bbclass)
COMMON_LICENSE_DIR: this is the directory that holds all the common
generic license files. currently meta/files/common-licenses (set in
poky.conf)
TODO:
- We should verify the common-licenses. I stripped these from my Ubuntu
10.10 system.
- We should allow the capability of licenses on the rootfs, although the
resulting image created would be a lot larger.
- More common-licenses. I don't include bzip, zlib, ICS.... I should,
but that means tracking down a lot of licenses.
- General cleanup of licensing and standardization of names. We should
standardize on a naming convention. What's in licenses.conf should
match up with what is in the recipes which should match with what is
in common-licenses. Outside the scope of this though. See:
http://bugzilla.pokylinux.org/show_bug.cgi?id=650
We need to manipulate the stamps when removing WORKDIR to indicte that
tasks like compile or install can't just rerun. The most effective method
to do this is to convert the layout to match that which would have been the
case had the system been build from sstate packages.
For example, we'd task stamps like:
xxx-1.4.7-r3.do_compile
xxx-1.4.7-r3.do_configure
xxx-1.4.7-r3.do_fetch
xxx-1.4.7-r3.do_generate_toolchain_file
xxx-1.4.7-r3.do_install
xxx-1.4.7-r3.do_package.emenlow
xxx-1.4.7-r3.do_package_write
xxx-1.4.7-r3.do_package_write_ipk
xxx-1.4.7-r3.do_package_write_rpm
xxx-1.4.7-r3.do_patch
xxx-1.4.7-r3.do_populate_sysroot.emenlow
xxx-1.4.7-r3.do_setscene
xxx-1.4.7-r3.do_unpack
and after rm_work, we'd have stamps of:
xxx-1.4.7-r3.do_package_setscene.emenlow
xxx-1.4.7-r3.do_package_write_ipk_setscene
xxx-1.4.7-r3.do_package_write_rpm_setscene
xxx-1.4.7-r3.do_populate_sysroot_setscene.emenlow
We also need to handle stamps in the form xxx-1.4.7-r3.do_package.MACHINE.TASKHASH
as used by some signature generators.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This commit changes the sysroots path to be machine specific.
Changes includes:
1) STAGING_DIR_TARGET and STRAGING_DIR_HOST points to machine specific
paths.
2) task stamp files. Adding ${MACHINE} info into stamp files for
do_populate_sysroots and do_package tasks. Add a BB_STAMPTASK_BLACKLIST
to keep native, nativesdk, crosssdk, and cross-canadian stamp unchanged.
3) siteconfig path. Separate the site config path for different machines
to avoid one machine adopting the cache file of another machine.
4) sstate. Add machine name to sstate manifest file.
Change relocation code for sstate paths since sysroot is machine.
Keep native, nativesdk, crosssdk, and cross-canadian unchanged.
5) toolchain scripts. Change the environment path to point to machine
specific sysroots in toolchain scripts bbclass.
6) Relocate la files when populating to a different machine of the same
architecture.
7) Exclude STAGING_DIR_TARGET and STAGING_DIR_HOST parameter from sstate
siginfo since they contain ${MACHINE} information.
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Some steps (like gcc-cross-initial, gcc-cross-intermediate and eglibc-initial)
will install to new locations to avoid file overwriting.
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
By default the linux-yocto recipes operate on the current branch
and use it as a trigger to locate the description of a board. This
model works well when using the git repo outside of a build system
since the commands can be simply invoked and will do something
useful. However, it does mean that you can't have two BSPs that
differ only by configuration, building out of a single branch
in the repository.
This means that you must have many branches for very similar
BSPs. This model is still preferred, but having the choice of
branching strategies is better.
With this change we can have multiple BSPs using a single branch
with the preferred description being hinted from the build
system by passing the $machine value to updateme/configme.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
If build from sstate results, the ${D} will not be installed. In this
case the creation of ipk package will be skipped, which will cause
the build failure.
Fix the issue by removing the judgement of ${D} existence.
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Original we used absolute path in sysroot-destdir for both native and
target recipes. This commit changes target recipes to use relative path
which is same as the image directory.
[sgw: merged with libtool sysroot work]
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
libtool v2.4 --with-libtool-sysroot eliminates the need for
la mangling.
Based on Khem Raj's OE commits.
Signed-off-by: Scott Garman <scott.a.garman@intel.com>
We do not do it for native recipes, as the native compiler should
fall back to a prefix of /usr and not solely depend on the sysroot.
Otherwise we end up staging everything in the native sysroot before
we start to build target recipes.
Also remove la mangling code, which is no longer necessary.
Commit derived from Khem Raj's OE commits.
Signed-off-by: Scott Garman <scott.a.garman@intel.com>
LSB library checks will look for some symbols of qt4 libraries. Enable "accessibility" and "sm" in order to pass the LSB test.
Signed-off-by: Jingdong Lu<jingdong.lu@windriver.com>
When using python recipes imported from OE (e.g. python-cheetah) parsing only succeeds when python has already been built due to the PYTHON_DIR references.
This commit syncs the classes with OE to make it work, but keeps the *.pyo removal from yocto.
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
This will be used later on when updating the distutils classes
A small deviation from OE is the use of PYTHON_BASEVERSION instead of non-deterministic python code. See https://lists.yoctoproject.org/pipermail/poky/2011-January/002320.html for some background info
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
remove sanity check for SDK_ARCH=i686 due to recent fixes in cross toolchain areas, etc. the old issue no longer exists
Signed-off-by: Jessica Zhang <jessica.zhang@intel.com>
Fixes "multiple repeat" or "nothing to repeat" errors when parsing recipes
with '++' in the file name.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
If PATCHRESOLVE is set to "noop", then it isn't supposed to try and
resolve patches at all, and thus does not need TERMCMD. Therefore,
make sure Poky doesn't check for (and possibly fail at finding)
TERMCMD if it isn't needed in the first place.
Signed-off-by: Derek Buitenhuis <dbuitenh@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
To reduce the time on sanity testing, we remove variable SHARE_IMAGE and use
a new variable TEST_SERIALIZE in local.conf. It is by default set to 1. Poky
will copy and boot the to-be tested image for only once. It will not remove
or kill the image and test cases will be serialized executed against the same
image. If it is set to 0, image is always be copied for each cases, which takes
much time. I had a experiment that latest qemuppc sato only takes 7 minutes to
finish 9 sanity test cases, which takes more than 20 minutes before.
I also removed sanity case "boot" from sato/sdk/lsb because the other cases for
these targets already cover the check point of "boot".
Signed-off-by Jiajun Xu <jiajun.xu@intel.com>
Richard Purdie
Lianhao Lu
Kevin Tian
Joshua Lock
Mark Hatle
Qing He
Beth Flanagan
Yu Ke
Dongxiao Xu
Dexuan Cui
Bruce Ashfield
Saul Wold
Scott Garman
Jingdong Lu
Koen Kooi
Jessica Zhang
Paul Eggleton
Jiajun Xu