Age | Commit message (Collapse) | Author | Files |
|
Parallel image builds race in package_update_index_ipk, so locking is
needed and for a proper lock package_update_index_ipk needs to be a
proper task.
Acked-by: Tom Rini <trini@embeddedalley.com>
|
|
This reverts commit f1cc64d92cc4023fe8180dc0079d010eec0d71b0.
|
|
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
|
|
Signed-off-by: Chris Larson <clarson@mvista.com>
|
|
Signed-off-by: Chris Larson <clarson@mvista.com>
|
|
Signed-off-by: Chris Larson <clarson@mvista.com>
|
|
dependencies".
Fixes a bug pointed out where a missing rdepends of a package installed into
an image wasn't resulting in a failure in the do_rootfs task.
Signed-off-by: Chris Larson <clarson@mvista.com>
|
|
Set IMAGE_KEEPROOTFS to something other than the empty string to keep
tmp/rootfs around after image creation, otherwise it will be removed, as it's
not a fully functional filesystem, and causes much confusion among users as a
result.
Signed-off-by: Chris Larson <clarson@mvista.com>
Acked-by: Tom Rini <trini@embeddedalley.com>
Acked-by: Marcin Juszkiewicz <marcin@juszkiewicz.com.pl>
|
|
into org.openembedded.dev
|
|
|
|
|
|
|
|
This patch unbreaks current behaviour which was introduced by
MACHINE_KERNEL_PR variable.
As most of target machines do not use it they have PR with broken value
(set to "r0" instead of value in recipe). I took other way which makes
both types of users happy -- those with MACHINE_KERNEL_PR in use and
those without it.
By default we set M_K_PR to empty string instead of "r0" - this allows
to check is it set at all or not. If it is set then we set PR to this value.
Otherwise we ignore existance of that variable and use PR from recipe.
Signed-off-by: Marcin Juszkiewicz <marcin@juszkiewicz.com.pl>
|
|
|
|
|
|
This reverts commit ba434bd1dc9dc8ed61953a1579ea9e993e81cbb2.
Acked-by: Richard Purdie, Leon Woestenberg
|
|
in /usr/lib specifically for its bits
|
|
|
|
management; only add it to ${PN} in any case.
|
|
* this method will install all linguas in $IMAGE_LINGUAS for all packages installed into the rootfs
* currently opkg only, guarded with a check
|
|
Currently just checks for variables which the recipe can avoid defining, as
the values end up the same as the defaults. Formerly known as kergoth_sanity,
but we all know such a thing is mythical ;)
Signed-off-by: Chris Larson <clarson@mvista.com>
|
|
Signed-off-by: Ihar Hrachyshka <ihar.hrachyshka@gmail.com>
Acked-by: Tom Rini <trini@embeddedalley.com>
Acked-by: Denys Dmytriyenko <denis@denix.org>
|
|
Scrollkeeper generates its XML database for every package when
installing it. The problem is that while building it sees only specific
package scrollkeeper stuff so we get wrong scrollkeeper xml file
generated for every gnome.bbclass package. Including the file into every
gnome package results in package contents conflicts.
This patch removes a quick hack made by Koen with previous commit.
Signed-off-by: Ihar Hrachyshka <ihar.hrachyshka@gmail.com>
Acked-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Koen Kooi <koen@openembedded.org>
|
|
to TMPDIR
* freeze.inc would work as well, but I got this working first
|
|
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
|
|
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
|
|
Detect if we are behind proxy by checking for HTTP_PROXY variable
and use specified proxy for http traffic.
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
|
|
the fdo mime data it overlaps between packages.
* better solution is needed for people wanting docs on device, but this makes packages at least installable
|
|
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
|
|
|
|
Adds a base_path_out convenience function, which prepares a full path for
display to the user. The initial implementation just makes it relative to
${TOPDIR}. This function is then used for some messages outputted to the
user (packaged-staging, patch application, clean, unpack tasks).
Signed-off-by: Chris Larson <clarson@mvista.com>
|
|
OpenEmbedded outputs a lot of messages that the user is likely to never
care about. We should only output something when it reflects upon their
recipe (i.e. unpacking their sources, applying their patches), or is quite
significant or unusual.
Signed-off-by: Chris Larson <clarson@mvista.com>
|
|
|
|
Hi,
I have access to a cluster of Itaniums that I want to use as
build machines for openembedded. This patch enables Itanium as a
build environment.
Signed-off-by: Marcin Juszkiewicz <hrw@openembedded.org>
|
|
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
|
|
apt ignores Packages.bz2 unless bzip2 exists in /bin/bzip2 (!!).
If it's in /usr/bin/bzip2 (e.g. SuSE build host) the image fails to build.
Also, remove commented-out check for DEPLOY_KEEP_PACKAGES.
Signed-off-by: Michael Smith <msmith@cbnco.com>
|
|
|
|
Comment on what this function is used for primarily. The default
gcc3-native that we build is called 'gcc-3.4.6' so look for that one
first.
|
|
machine-specific kernels.
|
|
slugos-extd image.
|
|
Gives us guaranteed sane GNU coreutils on any host. Should greatly ease
support for various non-GNU host systems and allow to safely remove some other
native packages like 'install-native'.
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
SDK packages.
We create package_qa_check_dev/package_qa_check_dbg checks. While in here
clarify the comments as to what both are doing. Since SDK packages don't
get split like normal ones, it's OK to have -dev stuff in non-dev packages.
Suggested by Richard Purdie
Acked-by: Koen Kooi <koen@openembedded.org>
Signed-off-by: Tom Rini <trini@embeddedalley.com>
|
|
From: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Otavio Salvador <otavio@debian.org>
Acked-by: Jeremy Lainé <jeremy.laine@bolloretelecom.eu>
|
|
Allows parallel build for different images.
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Signed-off-by: Koen Kooi <koen@openembedded.org>
|
|
IMAGE_VARS usage forces images to have init, dev management, login
management and other stuff. In cases where you don't need any of those
(like initramfs images or other special-purpose images) you need to
override 4 variables in order to get clean image that only installs
things from IMAGE_INSTALL.
So, this changes IMAGE_VARS to more meaningful IMAGE_BOOT and makes it
possible to build images without boot components with single
IMAGE_BOOT="".
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Koen Kooi <koen@openembedded.org>
|
|
When a package contains two files like "name*" and "name*smth" OE can't
properly package those, looping infinetely trying to build file list. The
reason is that first file name globs to two files and OE adds them both
into the list, then globs the first one again and gets two same files, etc.
Fix that.
Signed-off-by: Koen Kooi <koen@openembedded.org>
|
|
Places do_rootfs function before do_install in image creation process.
This change allows multiimage builds, when one image depends on another (before
that second image creation raced with do_rootfs for the first one) and uses
it for some purpose.
Multiimage builds are useful in situations like creating an installer image
with initramfs (built in OE) and an image of the target FS (also built in OE)
or for building virtualized system images to be placed on target FS.
Signed-off-by: Koen Kooi <koen@openembedded.org>
|
|
This makes it possible for overlays to have checksum.ini.
Signed-off-by: Petri Lehtinen <petri.lehtinen@inoi.fi>
Acked-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Denys Dmytriyenko <denis@denix.org>
Acked-by: Chris Larson <clarson@kergoth.com>
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
|