]> code.ossystems Code Review - openembedded-core.git/log
openembedded-core.git
7 years agooe-selftest: add a test for failing package post-installation scriptlets
Alexander Kanavin [Mon, 12 Mar 2018 16:49:43 +0000 (18:49 +0200)]
oe-selftest: add a test for failing package post-installation scriptlets

The test runs a scriptlet that has an intentionally failing command in the middle
and checks for two things:
1) that bitbake does warn the user about the failure
2) that scriptlet execution stops at that point.

The test is run for all three package types: rpm, deb, ipk.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agometa/lib/oe/package_manager.py: warn about failing scriptlets for all package types
Alexander Kanavin [Mon, 12 Mar 2018 16:49:42 +0000 (18:49 +0200)]
meta/lib/oe/package_manager.py: warn about failing scriptlets for all package types

Previously this was done only for rpm packages; now also ipk/deb scriptlet
failures are reported.

In the future this will become a hard error, but it can't yet happen
due to the legacy 'exit 1' way of deferring scriptlet execution to first boot which
needs a deprecation period.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agopackage.bbclass: run pre/post installation/removal scriptlets using sh -e
Alexander Kanavin [Mon, 12 Mar 2018 16:49:41 +0000 (18:49 +0200)]
package.bbclass: run pre/post installation/removal scriptlets using sh -e

This allows catching errors in the scriptlets which would otherwise
go unnoticed, e.g. this sequence:
====
bogus_command
proper_command
====
would work just fine without any visible warnings or errors.

This was previously done only for rpm packages; this patch replaces
the rpm-specific tweak with one that works for all package types.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agounfs3: Fix libtirpc usage for unfs3-native version
Richard Purdie [Mon, 12 Mar 2018 22:23:53 +0000 (15:23 -0700)]
unfs3: Fix libtirpc usage for unfs3-native version

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agolibtirpc: Extend to native and nativesdk recipes
Khem Raj [Mon, 12 Mar 2018 04:40:52 +0000 (21:40 -0700)]
libtirpc: Extend to native and nativesdk recipes

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agounfs3: Fix build with musl
Khem Raj [Mon, 12 Mar 2018 04:40:51 +0000 (21:40 -0700)]
unfs3: Fix build with musl

Should also fix build on new build hosts where
with glibc 2.27 rpc support is dropped in favor
of libtirpc

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agoopenssl_1.0.2n: improve reproducibility
Juro Bystricky [Sat, 10 Mar 2018 19:27:29 +0000 (11:27 -0800)]
openssl_1.0.2n: improve reproducibility

Improve reproducible build of:

openssl-staticdev
openssl-dbg
libcrypto

There are two main causes that prevent reproducible build, both related to
the generated file "buildinf.h":

1. "buildinf.h" contains build host CFLAGS, containing various build
   host references.  We need to pass sanitized CFLAGS to the script
   generating this file ("mkbuildinf.pl". )

2. We also need to modify the script "mkbuildinf.pl" itsel in order to
   generate a build timestamp based on SOURCE_DATE_EPOCH, if present in
   the environment.

Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agodbus-test-ptest: improve reproducibility
Juro Bystricky [Sat, 10 Mar 2018 18:58:45 +0000 (10:58 -0800)]
dbus-test-ptest: improve reproducibility

Remove build host references from additional files.

Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agoreproducible_build_simple.bbclass: simple environment for reproducible binaries
Juro Bystricky [Sat, 10 Mar 2018 18:57:49 +0000 (10:57 -0800)]
reproducible_build_simple.bbclass: simple environment for reproducible binaries

Export environmental variables needed for binary reproducibility with consistent values.

This class can be used either directly via:
INHERIT += "reproducible_build_simple"

or can be inherited by a more complex/complete bbclass, for example a bblass which
will crack SOURCE_DATE_EPOCH for each recipe.

Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agoclasses/recipes: Use expanded BUILD_REPRODUCIBLE_BINARIES value
Juro Bystricky [Sat, 10 Mar 2018 18:56:25 +0000 (10:56 -0800)]
classes/recipes: Use expanded BUILD_REPRODUCIBLE_BINARIES value

Replace the occurences of BUILD_REPRODUCIBLE_BINARIES with expanded
values ${BUILD_REPRODUCIBLE_BINARIES} so the variable does not need to be
exported.

Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agoyocto-uninative: Upgrade to 1.8 version with glibc 2.27
Richard Purdie [Sat, 10 Mar 2018 03:46:00 +0000 (19:46 -0800)]
yocto-uninative: Upgrade to 1.8 version with glibc 2.27

Now distros are starting to ship glibc 2.27 we need a uninatve version
which contains glibc 2.27 which is in the 1.8 version.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agolsbinitscripts: update to 9.79
Alexander Kanavin [Thu, 1 Mar 2018 14:00:42 +0000 (16:00 +0200)]
lsbinitscripts: update to 9.79

Switch to github as pkgs.fedoraproject.org is down.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agomklibs-native: refresh patches
Alexander Kanavin [Fri, 9 Mar 2018 18:56:14 +0000 (20:56 +0200)]
mklibs-native: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoattr: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:13 +0000 (20:56 +0200)]
attr: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibunwind: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:12 +0000 (20:56 +0200)]
libunwind: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibtiff: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:11 +0000 (20:56 +0200)]
libtiff: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agocryptodev: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:10 +0000 (20:56 +0200)]
cryptodev: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoblktrace: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:09 +0000 (20:56 +0200)]
blktrace: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoxorg-xserver: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:08 +0000 (20:56 +0200)]
xorg-xserver: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoacpica: remove unnecesary no-werror.patch
Alexander Kanavin [Fri, 9 Mar 2018 18:56:07 +0000 (20:56 +0200)]
acpica: remove unnecesary no-werror.patch

It became out of date (missing newly added files), and seems no longer necessary for builds.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agomesa: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:06 +0000 (20:56 +0200)]
mesa: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agodrm: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:05 +0000 (20:56 +0200)]
drm: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agogtk-doc: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:04 +0000 (20:56 +0200)]
gtk-doc: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agognome: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:03 +0000 (20:56 +0200)]
gnome: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agosysstat: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:02 +0000 (20:56 +0200)]
sysstat: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agorpcbind: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:01 +0000 (20:56 +0200)]
rpcbind: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agonewt: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:56:00 +0000 (20:56 +0200)]
newt: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agonet-tools: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:59 +0000 (20:55 +0200)]
net-tools: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoltp: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:58 +0000 (20:55 +0200)]
ltp: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoethtool: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:57 +0000 (20:55 +0200)]
ethtool: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agocups: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:56 +0000 (20:55 +0200)]
cups: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agotcf-agent: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:55 +0000 (20:55 +0200)]
tcf-agent: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoruby: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:54 +0000 (20:55 +0200)]
ruby: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoqemu: drop already applied glibc-2.25.patch
Alexander Kanavin [Fri, 9 Mar 2018 18:55:53 +0000 (20:55 +0200)]
qemu: drop already applied glibc-2.25.patch

Due to patch fuzz it was applied again in a different place.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoqemu: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:52 +0000 (20:55 +0200)]
qemu: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agopython: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:51 +0000 (20:55 +0200)]
python: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoopkg-utils: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:50 +0000 (20:55 +0200)]
opkg-utils: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agom4: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:49 +0000 (20:55 +0200)]
m4: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoi2c-tools: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:48 +0000 (20:55 +0200)]
i2c-tools: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agogcc: drop patch that is already upstream
Alexander Kanavin [Fri, 9 Mar 2018 18:55:47 +0000 (20:55 +0200)]
gcc: drop patch that is already upstream

Due to patch fuzz, it was applied again, so the same code sequence was
repeated twice. Not sure if that caused any bugs, but certainly wasn't
the right thing to do.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoe2fsprogs: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:46 +0000 (20:55 +0200)]
e2fsprogs: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoopenssl: drop openssl-1.0.2a-x32-asm.patch
Alexander Kanavin [Fri, 9 Mar 2018 18:55:45 +0000 (20:55 +0200)]
openssl: drop openssl-1.0.2a-x32-asm.patch

The patch was applied in a completely incorrect spot (due to fuzz),
no one noticed or complained. Meanwhile upstream says the issue
has been resolved differently:
https://rt.openssl.org/Ticket/Display.html?id=3759&user=guest&pass=guest

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoopenssl: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:44 +0000 (20:55 +0200)]
openssl: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoiproute2: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:43 +0000 (20:55 +0200)]
iproute2: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoavahi: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:42 +0000 (20:55 +0200)]
avahi: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agou-boot: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:41 +0000 (20:55 +0200)]
u-boot: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agopcmciautils: drop duplicate chunk from pcmciautils-018/makefile_race.patch
Alexander Kanavin [Fri, 9 Mar 2018 18:55:40 +0000 (20:55 +0200)]
pcmciautils: drop duplicate chunk from pcmciautils-018/makefile_race.patch

The new rule was patched into the makefile twice.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agopciutils: refresh patches
Ross Burton [Fri, 9 Mar 2018 18:55:39 +0000 (20:55 +0200)]
pciutils: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agodevtool: add --force-patch-refresh to 'modify' and 'finish' commands
Alexander Kanavin [Fri, 9 Mar 2018 12:27:47 +0000 (14:27 +0200)]
devtool: add --force-patch-refresh to 'modify' and 'finish' commands

This is very useful for updating patch context so that any fuzz is eliminated.
Simply issue:

devtool modify <recipe>
devtool finish --force-patch-refresh <recipe> <layer_path>

Without this flag, devtool will not deem the commits in the workspace
different to patches in the layer, even if the commits have different,
up-to-date context line in them.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agometa/lib/oe/patch.py: do not leave .orig files if a patch isn't perfectly matching
Alexander Kanavin [Fri, 9 Mar 2018 12:27:46 +0000 (14:27 +0200)]
meta/lib/oe/patch.py: do not leave .orig files if a patch isn't perfectly matching

Particularly, this was causing 'devtool modify' to erroneously add those
.orig files into commits. This was getting in the way, if the goal
was to amend/update those existing patches.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoliburcu: Explicitly add pthread options to cflags
Khem Raj [Fri, 9 Mar 2018 20:29:06 +0000 (12:29 -0800)]
liburcu: Explicitly add pthread options to cflags

Some architectures e.g. riscv gcc does not add -D_REENTRANT
when enabling pthreads. Help it here by adding these options
while gcc gets fixed

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agorecipes: Disable lttng for riscv
Khem Raj [Fri, 9 Mar 2018 11:00:49 +0000 (03:00 -0800)]
recipes: Disable lttng for riscv

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agopackagegroup-core-sdk: Disable SANITIZERS for riscv64
Khem Raj [Fri, 9 Mar 2018 11:00:47 +0000 (03:00 -0800)]
packagegroup-core-sdk: Disable SANITIZERS for riscv64

Dont build yet

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoe2fsprogs: Add comment on why touch is needed uninative-1.8
Richard Purdie [Fri, 9 Mar 2018 19:09:36 +0000 (11:09 -0800)]
e2fsprogs: Add comment on why touch is needed

Commit b32f3b655189fd89dcfce084b6fda0d379300f75 added this code
but we could do with a commit so people realise why its there.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agoperl: remove perl-enable-gdbm.patch
Alexander Kanavin [Thu, 8 Mar 2018 18:18:04 +0000 (20:18 +0200)]
perl: remove perl-enable-gdbm.patch

The change was already present in upstream, so we just applied it
again (see bug 10450 for why).

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoperl: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:18:03 +0000 (20:18 +0200)]
perl: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agopython: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:18:02 +0000 (20:18 +0200)]
python: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agopython-numpy: update to 1.14.1
Alexander Kanavin [Thu, 8 Mar 2018 18:18:01 +0000 (20:18 +0200)]
python-numpy: update to 1.14.1

Drop backported 0001-BUG-fix-infinite-loop-when-creating-np.pad-on-an-emp.patch.

Drop 0001-BUG-fix-infinite-loop-when-creating-np.pad-on-an-emp.patch as
upstream is using os.path.basename() instead now.

License-Update: License.txt file was update to list licenses of individual components;
not all of them are 3-clause BSD.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoirda-utils: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:18:00 +0000 (20:18 +0200)]
irda-utils: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agozlib: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:59 +0000 (20:17 +0200)]
zlib: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoutil-linux: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:58 +0000 (20:17 +0200)]
util-linux: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoppp: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:57 +0000 (20:17 +0200)]
ppp: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agosyslinux: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:56 +0000 (20:17 +0200)]
syslinux: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agomtd-utils: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:55 +0000 (20:17 +0200)]
mtd-utils: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agointltool: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:54 +0000 (20:17 +0200)]
intltool: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoautomake: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:53 +0000 (20:17 +0200)]
automake: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoapt: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:52 +0000 (20:17 +0200)]
apt: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoiptables: drop unnecessary patches
Alexander Kanavin [Thu, 8 Mar 2018 18:17:51 +0000 (20:17 +0200)]
iptables: drop unnecessary patches

These were adding definitions for the second time
(see bug #10450 for why) or adding an include that isn't anymore
necessary for musl builds.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agotcp-wrappers: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:50 +0000 (20:17 +0200)]
tcp-wrappers: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoparted: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:49 +0000 (20:17 +0200)]
parted: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibpam: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:48 +0000 (20:17 +0200)]
libpam: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agomdadm: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:47 +0000 (20:17 +0200)]
mdadm: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibidn: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:46 +0000 (20:17 +0200)]
libidn: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoghostscript: drop incorrectly applied patch
Alexander Kanavin [Thu, 8 Mar 2018 18:17:45 +0000 (20:17 +0200)]
ghostscript: drop incorrectly applied patch

The patch was adding a change to the source file that was already there,
so the lines of code were repeated twice. This didn't create a bug or a
security issue, but it may well have.

Long story:
https://bugzilla.yoctoproject.org/show_bug.cgi?id=10450

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoghostscript: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:44 +0000 (20:17 +0200)]
ghostscript: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoethtool: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:43 +0000 (20:17 +0200)]
ethtool: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agobash: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:42 +0000 (20:17 +0200)]
bash: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoat: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:41 +0000 (20:17 +0200)]
at: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoxset: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:40 +0000 (20:17 +0200)]
xset: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoclutter: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:39 +0000 (20:17 +0200)]
clutter: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agowayland: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:38 +0000 (20:17 +0200)]
wayland: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolttng-ust: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:37 +0000 (20:17 +0200)]
lttng-ust: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agodtc: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:36 +0000 (20:17 +0200)]
dtc: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agokmod: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:35 +0000 (20:17 +0200)]
kmod: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agogstreamer: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:34 +0000 (20:17 +0200)]
gstreamer: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibtiff: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:33 +0000 (20:17 +0200)]
libtiff: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibksba: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:32 +0000 (20:17 +0200)]
libksba: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibffi: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:31 +0000 (20:17 +0200)]
libffi: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agognutls: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:30 +0000 (20:17 +0200)]
gnutls: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agogmp: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:29 +0000 (20:17 +0200)]
gmp: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibical: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:28 +0000 (20:17 +0200)]
libical: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agonspr: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:27 +0000 (20:17 +0200)]
nspr: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoserf: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:26 +0000 (20:17 +0200)]
serf: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoshared-mime-info: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:25 +0000 (20:17 +0200)]
shared-mime-info: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agogrub: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:24 +0000 (20:17 +0200)]
grub: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoexpat: refresh patches
Ross Burton [Thu, 8 Mar 2018 18:17:23 +0000 (20:17 +0200)]
expat: refresh patches

The patch tool will apply patches by default with "fuzz", which is where if the
hunk context isn't present but what is there is close enough, it will force the
patch in.

Whilst this is useful when there's just whitespace changes, when applied to
source it is possible for a patch applied with fuzz to produce broken code which
still compiles (see #10450).  This is obviously bad.

We'd like to eventually have do_patch() rejecting any fuzz on these grounds. For
that to be realistic the existing patches with fuzz need to be rebased and
reviewed.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agokbd: avoid conflict with busybox
Pascal Bach [Thu, 8 Mar 2018 14:52:47 +0000 (15:52 +0100)]
kbd: avoid conflict with busybox

showkey can also be provided by busybox

Signed-off-by: Pascal Bach <pascal.bach@siemens.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agokernel: make copying of crtsavres.o conditional
Bruce Ashfield [Tue, 6 Mar 2018 18:11:17 +0000 (13:11 -0500)]
kernel: make copying of crtsavres.o conditional

As of the 4.13 kernel, there are configuration + linker combinations
that do not need (or build) crtsavres.o for ppc64 targets. The commit
of interest is:

    commit efe0160cfd40a99c052a00e174787c1f4158a9cd
    Author: Nicholas Piggin <npiggin@gmail.com>
    Date:   Fri May 12 01:56:52 2017 +1000

        powerpc/64: Linker on-demand sfpr functions for modules

        For final link, the powerpc64 linker generates fpr save/restore
        functions on-demand, placing them in the .sfpr section. Starting with
        binutils 2.25, these can be provided for non-final links with
        --save-restore-funcs. Use that where possible for module links.

        This saves about 200 bytes per module (~60kB) on powernv defconfig
        build.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
As such, our arch test for crtsavres.o is not enough, we add a secondary
existence check before trying the copy.

[YOCTO #12576]

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agopython3: Fix install purelib to make pip3-python work properly
Jason Wessel [Wed, 7 Mar 2018 20:00:37 +0000 (12:00 -0800)]
python3: Fix install purelib to make pip3-python work properly

The oe-core version of python3 patches the purelib use directory to
the system libdir so as to make it work with multilibs properly inside
the patch fix_for_using_different_libdir.patch with:

-        'purelib': '{base}/lib/python{py_version_short}/site-packages',
+        'purelib': '{base}/'+sys.lib+'/python{py_version_short}/site-packages',

The problem is that this broke the pip3-python package because the
install directory is out of sync when using a multilib version of
python.  When ever a module is installed with pip3 install that is a
purelib it will get installed to a location that python3 will never
reference and cause random failures.

This patch fixes the purelib install directory to match the purelib
use directory for externally managed python modules when using
multilibs.

Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agopackage_manager.py: Explicit complementary fail
Niko Mauno [Wed, 28 Feb 2018 16:20:08 +0000 (18:20 +0200)]
package_manager.py: Explicit complementary fail

When running bitbake -c populate_sdk <image_name>, it is expected that
packages matching SDKIMAGE_INSTALL_COMPLEMENTARY name mask (unless
declared in PACKAGE_EXCLUDE_COMPLEMENTARY) are installed to resulting
SDK. Underlying mechanism issues a package manager install call for set
of complementary packages. However the mechanism doesn't seem to inform
the user all too obviously in case the package manager command behind
install_complementary() method fails -- and since it is combined with
attempt_only=True option, user might end up wondering why several *-dev,
*-dbg packages are missing from resulting SDK.

Improve associated install() method behaviour in affected OpkgPM and
DpkgPM classes so that a problematic state of affairs becomes directly
obvious for bitbake user, resulting in shell output like:

  WARNING: someimage-1.0-r0 do_populate_sdk: Unable to install packages.
  Command '...' returned 1:
  Collected errors:
   * Solver encountered 1 problem(s):
   * Problem 1/1:
   *   - package somepkg-dev-1.0-r0.x86 requires somepkg = 1.0-r0, but
         none of the providers can be installed
   *
   * Solution 1:
   *   - allow deinstallation of someotherpkg-1.1-r1.x86

   *   - do not ask to install a package providing somepkg-dev

   * Solution 2:
   *   - do not ask to install a package providing somepkg-dev

Signed-off-by: Niko Mauno <niko.mauno@vaisala.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>