We need to iterate over the dictionary pulling out the values, not take the top
level keys. If we don't do this, we end up with dependencies on the values of
PACKAGES, not library dependencies.
(From OE-Core rev: 7219bca11f554fbe2ed30f1537491987d65e9316)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Without this, circular dependencies are found when attempting to build
multilib versions of qt4 (or bitbake world in a multilib enabled build).
(From OE-Core rev: b2e8cc5ae227656211fb7f32260e7dc4e2fb556e)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If CCACHE is in the whitelist then CCACHE_DISABLE probably should be too.
(From OE-Core rev: c03f76160e3cc3cb4fbf3cee114665c34bff06e6)
Signed-off-by: Mike Crowe <mac@mcrowe.com>
Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* sofar only formal changes, but to test that everything still builds the same
(From OE-Core rev: 054a0e6c850f92c03fbb6314702de4e6318ccd25)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* in preparation for upgrade to 4.8.3
(From OE-Core rev: 24bf06bbcda4c0dfdad33cdff6394faa52657bb9)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* faster lookup in FILESPATH as small bonus
(From OE-Core rev: 8cc54d2d154b2ed9f931da39d75dc9c135f5e26d)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* make it easier to override them in bbappend
* convert pulseaudio to more common -pulseaudio/-no-pulseaudio form
(From OE-Core rev: 34e3687394c6fa18ef0443d63b8d7d0a68c441e0)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* releases.qt-project.org has 4.8.1 as well as 4.8.3
(From OE-Core rev: f12df439b893c70a8cd271ff8b8e6d760b78a2b3)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This causes a build loop, when DRM depends on Cairo depends on Mesa
depends on DRM. We can safely remove it as it's only one libdrm example
program which uses Cairo, which we won't be needing. At least it's not
worth the build loop.
(From OE-Core rev: a6d305261dc925210185d8b70fb1a923e012153b)
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
binutils will build differently if this feature is enabled, so
make the do_configure step depend on it
(From OE-Core rev: 0788cf349fe37ef4a36c626dbc396c97d1ab14d7)
Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fixes these sorts of issues present on older gcc (CentOS 5.x in this case)
| cc1: error: unrecognized command line option "-Werror=implicit"
| cc1: error: unrecognized command line option "-Werror=nonnull"
| cc1: error: unrecognized command line option "-Werror=init-self"
| cc1: error: unrecognized command line option "-Werror=main"
| cc1: error: unrecognized command line option "-Werror=missing-braces"
| cc1: error: unrecognized command line option "-Werror=sequence-point"
| cc1: error: unrecognized command line option "-Werror=return-type"
| cc1: error: unrecognized command line option "-Werror=trigraphs"
| cc1: error: unrecognized command line option "-Werror=array-bounds"
| cc1: error: unrecognized command line option "-Werror=write-strings"
| cc1: error: unrecognized command line option "-Werror=address"
| cc1: error: unrecognized command line option "-Werror=int-to-pointer-cast"
| cc1: error: unrecognized command line option "-Werror=pointer-to-int-cast"
Also fixes:
makekeys-makekeys.o: In function `main':
makekeys.c:(.text+0x85): undefined reference to `__isoc99_sscanf'
makekeys.c:(.text+0xa7): undefined reference to `__isoc99_sscanf'
collect2: ld returned 1 exit status
make: *** [makekeys] Error 1
Older libc do not have this defined, we can use the -D_GNU_SOURCE
to the compiler to prevent generating calls to this function and
make linking work
(From OE-Core rev: 83c560ae282c1a28fd2c311c66debd02a69f1678)
Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This had nowhere near enough testing...
This reverts commit ffb6928f57.
(From OE-Core rev: f162f0ecc96fdfb564aad968e5b8bc670640ea68)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Versions earlier than 0.14 can't cope with 32 bit binaries on a 64 bit
system and vice versa. This results in problems for certain SDKMACHINE
combinations on certain hosts. By ensuring we build
chrpath-replacement-native we avoid this problems and the binaries work
correctly.
[YOCTO #3161]
[YOCTO #3201]
(From OE-Core rev: f89bced26de055817100d0b0e03094b031fcfd48)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
chrpath is assumed to be provided by the build host system. This means
we need to provide a replacement version and install into a specific directory
to avoid races.
(From OE-Core rev: 147c44c882a3dc667c079165d3ddc9d78b702286)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We assume chrpath is provided natively so it should be listed in ASSUME_PROVIDED.
(From OE-Core rev: 97a3ea712003e8d48dc68c282e656591f39d2d1a)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This converts the option to maintain the existing behaviour unless the option is
specified. We do specify the option during the builds themselves to ensure what
the users expects is built.
(From OE-Core rev: 0cc479699fe885049625d54c712b500c1b719e75)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fixes:
WARNING: Unable to get checksum for libart-lgpl SRC_URI entry art_config.h: file could not be found
which otherwise happens during parsing, even if libart-lgpl isn't being built.
(From OE-Core rev: a9245e0d1dd1bee4ac01fe0c73d95179033ba979)
Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
All variants of opkg are currently configured --disable-curl so there
seems no point in depending on it.
(From OE-Core rev: 875f1eb876c17c038a77bc7b7a5fed775d9fd3ea)
Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
libpcap uses libnl on Linux to support sniffing mac80211 devices, which could be
useful.
(From OE-Core rev: 052a8406e66c9dcccc1fc506a32cc1706b93467b)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Galago has been replaced with Telepathy and Folks, and has been unmaintained for
years.
(From OE-Core rev: b878a9b1af63603779850f420fc0feb2ef78da9f)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The test for network error in sanity.bbclass was negated.
(From OE-Core rev: 9fcd0866f0e30a50182434f6bcae13bf9575807f)
Signed-off-by: Bogdan Marinescu <bogdan.a.marinescu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The bb.utils.explode_dep_versions function has issues where dependency information
can be lost. The API doesn't support maintaining the correct information so this
changes to use a new function which correctly handles the data.
This patch also fixes various points in the code to ensure that we do not have any
duplicates in things that use explode_dep_versions.
A new sanity test to test the contents of the R* variables is also added.
[Some changes from Mark Hatle <mark.hatle@windriver.com>]
(From OE-Core rev: 16a892431d0c0d03f8b561b92909cf2f11af4918)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
People keep copying this code and its confusing and unnecessary. Remove the
bad examples to try and stop this happening.
(From OE-Core rev: 48aa4b00cfb7f01195c6d20b7ba660715fe792ef)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Without this change, anything using linuxstdbase would incorrectly
try and pull in X dependencies.
(From OE-Core rev: 664fce61c9c49276039b18606e1e88585b47f724)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Even when installing onto a live target system, calling sync() during
package installation is of somewhat questionable benefit. But calling
it on the build host during rootfs construction is certainly useless and
can cause I/O to stall for several seconds on even a moderately sized
host which is clearly not desirable.
(From a patch originally by Mike Crowe.)
(From OE-Core rev: f48a68177510e8f2d4fcc3725a6dfc41a9a8e96b)
Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In deb and ipk, < means <=, while > means >=... there is a different
operator << and >> that means < and >, so we map them when constructing
the packages.
(From OE-Core rev: bbcc78d8ff03725ce5b3b65ce24025c3da45f2ab)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If the STAGING_DIR_KERNEL is set in the multilib.conf, then it may be
set incorrected. The evaluation happens before TMPDIR and LIBC are
defined in other components.
Moving the definition process to the multilib.bbclass ensures that
everything has been loaded before it is set.
(From OE-Core rev: 6bd87edc383b40e300b0ef4bf851c39b698305cd)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When building target packages, it used to be enought to check for PN == BPN, however
with the multilib configurations, this can lead to subtle errors. Change instances
of PN == BPN, to ${CLASSOVERRIDE} == 'class-target'.
(From OE-Core rev: acc988272b4e74a9ad1e6da5af5b2d208584197b)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In certain cases with BerkleyDB 5.3.x we are getting the error:
db3.c:1443: dbcursor->pget(-30999): BDB0063 DB_BUFFER_SMALL: User memory too small fo
See https://bugs.launchpad.net/rpm/+bug/934420 for more information.
It appears to be some type of a bug in the BerkleyDB 5.3.x. In an attempt
to workaround the problem, when we encounter this situation we attempt
to adjust the size of the mmap buffer until the call works, or we
end up trying 10 times. The new size is either the updated vp->size
from the failed pget call, or the previous size + 1024.
If DBI debugging is enabled, additional diagnostics are printed, otherwise
a basic retry and success message is added to show that the failure was
resolved.
(From OE-Core rev: bfb2906206158748d0be33baf7984cf885756da1)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Remove the optimzation append from recipe and add the patch that is in the rpm
cvs repo, http://www.mail-archive.com/rpm-cvs@rpm5.org/msg08907.html. The -O2
optimzation append is removed since it can limit debugging options that are
provided when -O0 is used.
This was tested by setting: SELECTED_OPTIMIZATION = "-O0"
(From OE-Core rev: d109c6bd163469d6281d20174e4b79cb63483cd4)
Signed-off-by: Morgan Little <morgan.little@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There is an issue that is caused when doing the install step of rpm on systems
with high parallelization where two jobs of make will fight for the same file
while installing the sub-directory lua. This is caused by the same makefile rule
being called twice in a way that both could be trying to install at the same
time.
This fix renames the linking rule so it will always be run after the needed
files are added and removed it's dependency so the required rule would only
run once.
This was tested heavily using ppss to run mutliple installs in parallel. This
wouldn't happen in practise but it was tested will all the individual rules as
well.
(From OE-Core rev: d05c5da6b972db97d3eb66b659f5641368c9ebe4)
Signed-off-by: Morgan Little <morgan.little@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When attempting to build with uuid and all tests compiling will error because
uuid.h doesn't exist in the rpm tarball. Fix this by changing the include to
use the one in ossp which solves the issue.
The recipe already depends on ossp so ossp/uuid.h will be there when rpm-native
is built.
(From OE-Core rev: 52ae2c2439bcb78323f61a3666e9b630b3a40b15)
Signed-off-by: Morgan Little <morgan.little@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Some items were listed multiple times in DEPENDS, avoid this situation.
Note, PR was not incremented as no change to the build process occurs.
(From OE-Core rev: e234af467eac7d0313fae3e87eb1b34725309bb5)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This likely requires some form of license to use in a commercial
product.
(From OE-Core rev: 1d41af288f3db07a5dc47436443daf95e243904f)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When merging the xserver-xorg fix the to use RDEPENDS in
xserver-xorg-module-exa the RCONFLICTS has not been removed by
mistake. This drops the RCONFLICTS to properly fix it.
(From OE-Core rev: d83e218dc480a09befddf8b934d774519cdbacb5)
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This ensures consistent build results and avoids build failures when compiler flags
change for example.
(From OE-Core rev: a5ff8396cad130f809f8f8da49bb38e6f80f923c)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We're going around in circles trying to fix the sed expressions and making one case
work and others not work. This patch fixes the base configuration file so we have
non-overlapping substitutions. I've tried to significantly clean up various problems
that were occurring once and for all.
This will hopefully resolve all the issues people have been seeing with incorrect perl
paths.
(From OE-Core rev: 31ff70794ecc431431476f81c8934fff25383613)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Also supports a new altivec TUNE_FEATURE
(From OE-Core rev: 4586c24ad156773568cd38794936b8af62e862be)
Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie
Mike Crowe
Martin Jansa
Daniel Stone
Matthew McClintock
Ross Burton
Phil Blundell
Bogdan Marinescu
Mark Hatle
Paul Eggleton
Otavio Salvador