| Age | Commit message (Collapse) | Author |
|---|
|
This solves a regression introduced by OE-Core:9b75f6a (kernel:
use oldnoconfig before yes '' | make oldconfig).
The original oe_runmake explicitly calls 'die' command in case of
failure so the fallback code never runs. The fallback code needs
to handle the oe_runmake return code to call the backward
compatible callback so we introduced a new command called
oe_runmake_call which is used by oe_runmake.
This does not change the functional behaviour of oe_runmake so it
avoids any change except for the code which does need to handle
the oe_runmake exit code.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Koen Kooi <koen.kooi@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Unpacking an src.rpm uses rpm2cpio.sh, which requires 'file'.
Without this, builds of rpm on a host without 'file' installed will fail with
very strange messages.
Signed-off-by: Christopher Larson <kergoth@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
gcc has cross and target components with a shared workdir. The unpack umask
settings need to match for all of these. We need to use strings in each
case to ensure the sstate code matches them correctly.
This patch tweaks various things to ensure the change adding the unpack umask
change doesn't break the compiler builds.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
* when git checkouts files from fetched clone it respects system umask
and creates files with different permissions, if such files are copied
to packages, resulting target images have also different permissions
on them.
* we need reproducible builds across different builders with different
system umask, so set 022 umask
[YOCTO #5590]
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
PREFERRED_PROVIDER_virtual_pkg has been incorrectly set with more
than one multilib prefixes. For example, if we have two alternative
multilibs lib64 and lib32, PREFERRED_PROVIDER_virtual_pkg will be
set to lib32-lib64-pkg or lib64-lib32-pkg, depending on which
multilib shows up first in the list.
Signed-off-by: Lei Liu <lei.liu2@windriver.com>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The PACKAGECONFIG flags were iterated over using dict.items(), but this
returns the items in an undefined order. As this order determines the
EXTRA_OECONF append order, we can get EXTRA_OECONF which are functionally
equivalent, but whose contents differ, resulting in not using shared state
archives we should be using.
Signed-off-by: Christopher Larson <kergoth@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
These have been deprecated for a long time, convert the remaining
references to the correct modules and prepare for removal of the
compatibility support from bitbake.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
base.bbclass had code which handled the PNBLACKLIST in case of
multilib use. This is better to be done in the blacklist.bbclass so it
has all logic in a single place.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
The build banner displayed at the start of builds can be misleading since
the data store has not been finalised. As easy way to illustrate this is
to use something like:
DEFAULTTUNE = "i586"
DEFAULTTUNE_<machineoverride> = "core2"
and the banner will display the i586 tune yet the core2 tune will be
used. We can avoid this if we finalise a copy of the data before
displaying it.
[YOCTO #4225]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Now that bitbake supports masking events for event handlers, lets use
this so event handlers are only called for events they care about. This
lets us simplify the code indentation a bit at least as well as mildly
improving the event handling performance.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
dependencies
The do_package_setscene task needs to depend on fakeroot in order to correctly
install its files.
We can whitelist the dependency in the sstate handling code for some
performance improvements since we only need this if we're installing the
package from sstate.
Also use an append operator in base.bbclass for clarity.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Update older code to use modern exception handling syntax which
is the form accepted by python 3.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The use of SOC_FAMILY here is old code and SOC_FAMILY is now implemented by
MACHINEOVERRIDES behind the scenes. It therefore makes more sense to use
the replacement value in this code. Just like SOC_FAMILY, this is a ":"
delimited variable so we should iterate over the components, not use
the value directly.
Finally, MACHINEOVERRIDES contains MACHINE so we don't need to check that
directly.
This makes the functionality match what most users would expect it to do
and is also compatible with the way things previously worked.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When processing the blacklists, we should avoid cross-canadian packages, as they
will not have any multilib prefixes to expand.
Similarly look for "virtual/nativesdk-" in addition to the existing "nativesdk-".
These items should also be ignored.
Finally, in order to avoid undeterministic variable key expansion, such as:
MYVAR = "foo"
PREFERRED_PROVIDER_${MYVAR} = "bar"
PREFERRED_PROVIDER_foo = "foobar"
during the multilib processing of PREFERRED_VERSION and PREFERRED_PROVIDER,
the code was changed to rename the variable key, to the final key. This along
with the existing code avoids the problems.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
Its generally useful for devshell to end up in the fakeroot environment. If
a user needs to exit it, PSEUDO_UNLOAD=1 <command> works, its usually
harder to enter the envionment.
[YOCTO #3374]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Currently, various standard library operations like os.walk(),
os.path.isdir() and os.path.islink() each call stat or lstat which
involves a syscall into the kernel. There is no caching since they could
conceivably have changed on disk. The result is that for something like
the do_package task of the kernel we're spending over two minutes making
868,000 individual stat calls for 23,000 files. This is suboptimal.
This patch adds lib/oe/cachedpath.py which are a set of replacement
functions for these operations which use cached stat data rather than
hitting the kernel each time. It gives a nice performance improvement
halving the build time of the kernel do_package.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
It's too easy to cause rebuilds because the DISTRO_FEATURES have changed in
meaningless ways (such as re-ordering or duplicate items). Help stop this by
checking if the feature to be back-filled is already present.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
git no longer supports the use of GIT_CONFIG which defeats the purpose of
GIT_CORE_CONFIG and the generate_git_config() function. Remove it.
Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
oe_runmake() tried to set MAKE a second time, remove it since this is
(nowadays?) redundant WRT the normal export MAKE= in the generated
scripts.
Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The cmake class uses EXTRA_OECMAKE to adjust the cmake configuration. This
patch adds support for this, if the cmake class is used by the recipe.
Signed-off-by: Philip Balister <philip@balister.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Backfilling DISTRO_FEATURES and MACHINE_FEATURES with _append statements happens
too late to use those variables with conditional inherits, like this:
inherit ${@base_contains('DISTRO_FEATURES','sysvinit','update-rc.d_real','',d)}
Instead, do the backfilling at ConfigParse time so that it happens earlier in
the parse, which results in that inherit behaving as expected when sysvinit was
backfilled.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Handle the potential case where the recipe level LICENSE contains an
incompatible license, but all of its emitted packages are in fact compatible.
Without this adjustment, it would skip such a recipe.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Spamming potentially over a screenful of messages due to their
INCOMPATIBLE_LICENSE selection seems a bit unnecessary. If the distro was
configured that way, the user likely won't care, and if the user set it, they
likely know the implications, so make the messages debug instead.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Allow INCOMPATIBLE_LICENSE to be a whitespace-separated list of
incompatible license strings and/or glob patterns.
Also fix wildcarding: the string in INCOMPATIBLE_LICENSE was clearly
intended to match with wildcards (e.g. "*GPLv3" to match both GPLv3
and LGPLv3), but this was broken because of a bug in return_spdx()
which would die with a runtime error when there was no SPDXLICENSEMAP
entry for the string.
Signed-off-by: Andy Ross <andy.ross@windriver.com>
Signed-off-by: Elizabeth Flanagan <elizabeth.flanagan@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
In the interests of simplifying things, remove P and PN from FILESPATH,
instead relying on the BP and BPN versions which work in 99% of cases.
In any problematic case such as a -native only recipe, either the patch
directory can be renamed or the recipe can set FILESPATH specifically.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Files are very rarely, if ever placed in ${PF}. If a recipe needs to do this,
it can easily append to FILESPATH so it makes sense to drop this from the
default search path.
Equally, using FILE_DIR as part of the search path leads to 'bad' SRC_URI
entries and/or file layouts which are not preferred. I'm therefore of the
opinion we should also remove this from FILESPATH and encourage people to
cleanup any places this breaks my correcting the layouts to match the standard
or worst case adding to FILESPATH in recipes that need it.
These changes work towards making the system more friendly as users won't be
greeted with huge search paths we rearely use making the "correct" layout
more obvious.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Crosssdk packages are not actually multilib packages, so treat them
the same as other nativesdk packages in the multilib, base, and
classextend components.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
Add support for the generation of cross-canadian packages.
Each cross-canadian package has:
PN = "pkg-cross-canadian-${TRANSLATED_TARGET_ARCH}"
in order for that to be evaluated properly with multilibs enabled, it was
necessary to detect both the presence of the cross-canadian packages and
then update the vars using the OVERRIDE for the multilib. Additional checks
were made to ensure that any dependency that sais "cross-canadian" did not
get prefixed with the MLPREFIX.
Also, make sure that even when building multilib cross-canadian packages,
we only use the single SDK PACKAGE_ARCH, we don't want or need variants.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
These variables correspond to the PACKAGE_ARCH list combined with the TARGET_VENDOR
and TARGET_OS values. These can be used to traverse the pkgdata structure.
Setting these once in base.bbclass stops pkgdata needing to recalculate the values
and is also useful for the reworked shlibs code in a patch that will follow this.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
As discussed on the mailing lists, using a suffix to package names is
hard and has lead to many recipes having to do PKGSUFFIX games. Its
looking extremely hard to scale nativesdk much further without hacking
many recipes.
By comparison, using a prefix like multilib does works much better and
doesn't involve "hacking" as many recipes. This change converts nativesdk
to use a prefix using the existing multilib infrastructure.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
It was pointed out we have a number of weird indentations in the python functions.
This patch cleans up 3, 7 and other weird indentations for the core bbclass files.
It also fixes some wierd (odd) shell function indentation which my searches picked up.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
PACKAGECONFIG flag code only handles that when it has 3 or 4 items in flag,
it may have a stale data if some flags doesn't need DEPENDS and RDEPENDS.
Signed-off-by: Yao Zhao <yao.zhao@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
It turns out that the result of getVarFlags is not a list, it's
a dict. So "getVarFlags(...) or []" does not reliably produce
something with a .items. This escaped detection because our
local build environment never ends up running builds without
PNBLACKLIST entries.
Signed-off-by: Peter Seebach <peter.seebach@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The PNBLACKLIST feature does not currently work with multilibs,
because they have different ${PN}. The obvious thing to do is to
do this at the same point that we do the PREFERRED_PROVIDER
and PREFERRED_VERSION fixups. (Making the PNBLACKLIST check
do the for-each-multilib check requires it to do the multilib
list generation repeatedly.)
Signed-off-by: Peter Seebach <peter.seebach@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
This is useful when you have a rather large set of compatible distros. For
example: Centos 5.4, 5.5, 5.6, etc, RHEL server 5.x, RHEL workstation 5.x.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
running on
This can then be used by the sstate code to mark native and cross packages
as being specific to a given distro.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
handling of native path issues
If something is listed in ASSUME_PROVIDED but also manually built (like bzip2-native and
bzip2-replacement-native), we need to ensure that the installed binaries are only accessed
for things which list the provider in DEPENDS.
This patch adds a generic mechanism to handle this instead of everything reinventing the
wheel. EXTRANATIVEPATH += "perl-native" will ensure that ${STAGING_BINDIR_NATIVE}/perl-native
is added to PATH.
This means that:
a) Such providers should install to ${bindir}/xxx-native
b) Should PROVIDE += "xxxx-replacement-native"
c) That users should DEPEND on xxx-replacement-native and have EXTRANATIVEPATH += "xxx-native"
This patch enables this at the core level, the bzip2 recipe still needs adjusting to work
like this along with adjustment of the users of bzip2-replacement-native (python-native?).
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The bb and os modules are always imported so having these extra import calls
are a waste of space/execution time. They also set a bad example for people
copy and pasting code so clean them up.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
|
|
exclusion list
This appears to be an oversight in the original implementation. All of the
host package types were being ignored except for the SDK cross-canadian type.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
|
|
Currently, ccache is used if it is present. When building from scratch it gives
no performance improvement and creates a ton of empty directories even when its
not in use.
This change moves ccache support to a bbclass file which the user can choose to
enable. This should make builds more determinstic and make it easier/clearer
to the end user when its being used and when it is not.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Connect the new fetcher file checksum code so that we get a list of the
files to be checksummed at parse time.
The file-checksums flag will not be read unless we are using a version
of BitBake that supports the function we call within it, so it is safe
to include this change even when the metadata will still be used with
older versions of BitBake.
Implements [YOCTO #2044].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
|
|
Signed-off-by: Christopher Larson <kergoth@gmail.com>
|
Otavio Salvador
Koen Kooi
Christopher Larson
Richard Purdie
Martin Jansa
Lei Liu
Mark Hatle
Ross Burton
Darren Hart
Bernhard Reutner-Fischer
Philip Balister
Christopher Larson
Andy Ross
Yao Zhao
Peter Seebach
Paul Eggleton