Age | Commit message (Collapse) | Author |
|
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
* empty MACHINE_OVERRIDES adds :: to OVERRIDES (then in FILESPATH skips
every directory after this)
* strip() is needed to remove space in case some distribution has ie
MACHINE_OVERRIDES ?= ""
and then
MACHINE_OVERRIDES += ""
or
MACHINE_OVERRIDES ?= "${MACHINE_CLASS}", where MACHINE_CLASS is empty again
While FILESPATH was fixed separately, having :: in OVERRIDES still makes
bitbake 1.8.18 to throw a "RuntimeError: Set changed size during iteration"
exception when used together with collections.inc, since bitbake is being
re-spawned in there. More details here:
http://thread.gmane.org/gmane.comp.handhelds.openembedded/38254/focus=38595
Acked-by: Chase Maupin <chase.maupin@ti.com>
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
|
|
Note that this change requires a version of BitBake with the "git subpath"
patch applied.
Signed-off-by: Paul Eggleton <bluelightning@bluelightning.org>
|
|
- reverse OVERRIDES
- drop empty overrides
- uniq the list
- use nested variable references rather than getVar
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
Given the current implementation of OVERRIDES in bitbake, the variable is
expected to contain elements in the order least specific to most specific,
however, our current usage of it does not match that. As one example, "local"
is supposed to always be the most specific override, yet currently it's the
least specific. As another example, currently the target architecture is seen
as more specific than the machine, which is also clearly wrong.
Big thanks to Chase Maupin for investigating and identifying this long
standing issue.
It becomes clear that a reversal of the current value will bring us to a more
sane behavior, and avoids the need for the dual overrides hack mentioned in
the comments, so this implements this reversal, and drops the unnecessary and
confusing comments.
This also introduces a MACHINE_OVERRIDES variable as a generic mechanism to
inject overrides elements which are more specific than the distro but less
specific than the machine, which is where things like MACHINE_CLASS or
SOC_FAMILY or the like would go. This variable is *space* separated, to make
it easier and more convenient to assemble the variable incrementally.
Reported-by: Chase Maupin <chase.maupin@ti.com>
Signed-off-by: Chris Larson <chris_larson@mentor.com>
Acked-by: Chase Maupin <chase.maupin@ti.com>
Acked-by: Khem Raj <raj.khem@gmail.com>
|
|
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
By default, pkg-config falls back to checking the default search path for .pc
files, which is the native location. If you relocate it, it will be unable to
hit that location, which can cause problems in certain cases. In particular,
gnome-bluetooth requires gtk-doc-native to run its tools, and uses pkg-config
to determine its availability.
Fix by adding the native pkg-config locations to the PKG_CONFIG_PATH.
Signed-off-by: Chris Larson <chris_larson@mentor.com>
Acked-by: Khem Raj <raj.khem@gmail.com>
|
|
* Set default weak to "no"
* Use in TARGET_LDFLAGS
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Martin Jansa <Martin.Jansa@gmail.com>
Acked-by: Frans Meulenbroeks <fransmeulenbroeks@gmail.com>
Acked-by: Koen Kooi <k-kooi@ti.com>
|
|
libdir to be something else than '/lib'
* Add base_libdir_native as target libdir could be different e.g. lib32
lib64
* Dont assume lib to be system libdir during packaging.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Chris Larson <chris_larson@mentor.com>
|
|
Signed-off-by: Khem Raj <raj.khem@gmail.com>
|
|
Boards with small flash and large ram need all the compression they can get.
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
--accept
* --trust-server-cert is supported only in >=svn-1.6
http://subversion.apache.org/docs/release-notes/1.6.html
* --force allows to overwrite existing dirs
ie
matchbox-panel-2-icon-themes_0.0.1.bb
matchbox-panel-2_svn.bb
are checkouting/updating tiwo different SRCREVs and every few builds
you end up removing svn checkout from downloads dir, just because it
refuses to upgrade to newer revision (applets dir already exists there).
* --accept to overwrite any local modifications (as we want clean checkout)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
Fixes fetching from https:// svn repos with self-signed certs.
Signed-off-by: Roman I Khimov <khimov@altell.ru>
|
|
perl makes use of this variable and it doesn't match how we use it.
Since it doesn't need to be exposed like this, don't.
No acks, but "sounds good" to Koen and "don't see why not" from Roman
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
paths to new multi-machine friendly toolchain install location.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Koen Kooi <k-kooi@ti.com>
Acked-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
Hi all,
We recently noticed that apr and apr-util recipes were not fetching correctly.
It appears as though the APACHE_MIRROR variable in the bitbake.conf file.
Below is a trivial patch to update the variable.
Regards,
Sean
Signed-off-by: Sean Hudson <Sean_Hudson@Dell.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
|
|
Signed-off-by: Khem Raj <raj.khem@gmail.com>
|
|
sysroot
* Update references to CROSS_DIR in classes and bitbake.conf.
* This patch is based on Joshua Lock's work in poky.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Tom Rini <tom_rini@mentor.com>
Acked-by: Richard Purdie <rpurdie@rpsys.net>
Ackef-by: Chris Larson <clarson@kergoth.com>
Acked-by: Martin Jansa <Martin.Jansa@gmail.com>
Acked-by: Koen Kooi <koen@openembedded.org>
|
|
This moves the current CCACHE over to local.conf.sample and documents
when it's useful to enable.
Signed-off-by: Tom Rini <tom_rini@mentor.com>
|
|
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
It's unnecessary, and can cause problems with amend.bbclass.
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
These don't change, so their initial values should be just fine. Apparently,
in a typical build, those os.uname() snippets are called around 46k times,
which seems a tad excessive :)
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
${PN}-dbg
Some libraries provide gdb autoload script as a debugging aid.
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
Firstly adds tooling from Poky to scan packages and fix up paths in libtool
files, we then build on this to further post-process native packages (native,
cross, sdk) and fix up any references to the STAGING_DIR in non-binary files.
Secondly adds support for fetching packaged staging packages from a mirror by
setting PSTAGE_MIRROR to point to the root of a pstaging directory.
Signed-off-by: Joshua Lock <josh@linux.intel.com>
|
|
populate_staging task to populate_sysroot
This change, pulled from Poky, makes the purpose of the staging directory more
obvious and the taskname more true to what it now actually does.
The layout version number is increased due to the change in layout but code to
convert existing directories and insert a symlink for backwards compatibility
is included.
This patch also includes fixes for all recipes which reference the directory.
Signed-off-by: Joshua Lock <josh@linux.intel.com>
|
|
you're using AUTOREV
* For now disable autoincrement of revision counter in SRCPV, whoever
wants it, should enable it in local.conf or distro config.
(Set it empty BB_LOCALCOUNT_OVERRIDE = "")
* Revision counter is incremented only locally (bad for multiple
builders filling shared feeds), LOCALCOUNT can be used to maintain
counter globally in OE metadata (LOCALCOUNT value set in recipe).
* Make sure you disable it, if you want always upgradable paths for git
recipes without bumping LOCALCOUNT with every SRCREV change
or if you're using AUTOREV feature.
* LOCALCOUNT == value for revision counter in GIT recipes (NNNN in
PV="1.0+gitrNNNN+a7401084a72285c8d3485bacf4eed593f303b1b7-r4.4")
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
It can interfere with the bblayers stuff in bitbake master.
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
type, if not set to external
Signed-off-by: Koen Kooi <k-kooi@ti.com>
|
|
Makes referencing it easier.
Signed-off-by: Roman I Khimov <khimov@altell.ru>
Acked-by: Khem Raj <raj.khem@gmail.com>
|
|
Acked-by: Richard Purdie <rpurdie@linux.intel.com>
Acked-by: Marcin Juszkiewicz <marcin@juszkiewicz.com.pl>
Signed-off-by: Chris Larson <chris_larson@mentor.com>
|
|
URLs (from Poky)
|
|
* ETCDIR instead SYSCONFDIR, because there already was STAGING_ETCDIR_NATIVE
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
* initrd images need to be prepared with mkimage to be usable from u-boot
Signed-off-by: Steffen Sledz <sledz@dresearch.de>
Acked-by: Martyn Welch <martyn.welch@gefanuc.com>
Acked-by: Tom Rini <tom_rini@mentor.com>
|
|
When Java things are build for more then one architecture strange things
happen:
- first architecture builds fine
- next architectures fails to build
Thats due to lack of JAR files as they were built as PACKAGE_ARCH="all"
but staged as PACKAGE_ARCH="TARGET_ARCH" so BitBake do not build them
again but Java classes are unable to find them.
Solution is to stage JAR files to non-ARCH directory. I used
staging/java for it and build tested with stable/2009 and
org.openembedded.dev branches.
Signed-off-by: Marcin Juszkiewicz <marcin@buglabs.net>
|
|
angstrom, shr preferred providers
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
In OpenEmbedded we used squashfs 4.0 with external patch for LZMA
compression support. In the meantime, squashfs has mainlined LZMA
support in its CVS version.
This patches switches to the upstream squashfs-tools and modifies
bitbake.conf to match the slightly different arguments to mksquashfs.
Signed-off-by: Leon Woestenberg <leon@sidebranch.com>
|
|
This reverts commit 78ec619c698b227e0d0f310f49c4a932ce9f0db6.
Upstream squashfs-tools has now-official support for LZMA but
requires different arguments. This commit added the new
argument, but I later decided to unify this commit with the
actual recipes -- so that in case of problems, it can be traced
back to a single commit.
Signed-off-by: Leon Woestenberg <leon@sidebranch.com>
|
|
|
|
This reverts commit 2f982a8b546846d705bdb6e75a58692e224b6726.
|
|
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
|
|
(from Poky)
* Drop the inherit sdk since it only gains us a single path and a world of pain.
* Fix up paths to use SDKPATH instead of prefix.
* Drop layout_* variable usage
* Correct opkg control file locations
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
|
|
* Holger Hans Peter Freyther <zecke@openmoko.org> did for the OM git
tree that was never merged back into OE.
* Taken from xora/angstrom-srcpv branch where it rot for long enough
Signed-off-by: Graeme Gregory <dp@xora.org.uk>
Acked-by: Richard Purdie <rpurdie@linux.intel.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
fixing some multimachine corner cases
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
|
|
base_prune_suffix() function from Poky and extend the number of special suffixes
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
|
|
allow nextgen SDK generation (from Poky)
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
|
|
* add support for LZMA-compressed cpio images for use as Linux
RamDisk images.
|