If sed-native is built before these programs, hardcoded paths to sed-native
can end up in scripts and other parts of the system which may cause issues
if they are later used from sstate and sed-native is not installed.
To avoid this, this patch changes the global site configuration to
specify that plain "sed" is fine to be used. We need to spell this
out for gcc since it doesn't see the site files since we don't autoreconf
it. We can remove the values from libtool. We tell perl to use "/bin/sed"
since it requires a path and the system sed should be just fine for it.
[YOCTO #4971]
(From OE-Core rev: 2ec171cb188601bf18c6c2895870907024b1c52a)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In configure there is a runtime check to determine buggy getaddrinfo and as we
are cross-compiling this check goes wrong. If ipv6 is enabled in python, the
failed test is reported as fatal error with the message
Fatal: You must get working getaddrinfo() function.
Setting ac_cv_buggy_getaddrinfo=no deactivates the runtime check and allows to
compile python with ipv6 enabled.
This commit was derived from the commit 700b75e7661062aa93cf81205b78c8bf7609922d
in the Classic OpenEmbedded Development Tree.
(From OE-Core rev: 95cc8015a867235bed6d6cde08ab87f5351a66c7)
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@oss.bmw-carit.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(borrow from OpenEmbedded with below tweaks)
Enhance login_defs_pam.sed according to shadow source, to ensuer we don't
leave any unknown definitions in /etc/login.defs when pam is enabled
no need for --disable-account-tools-setuid which is detected upon pam
automatically, and no specific CFLAGS append
move shadow site options to generic site files
adjust indention
RDEPENDS on a list of pam-plugins since they're separately packaged
test with both pam enabled and pam disabled. when pam is enabled, tried
some same tweak with desired effect.
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
This reverts commit da9f079481 which turned out to
cause more problem that it fixed with libtool failures of:
| i586-poky-linux-libtool: link: unsupported hardcode properties
| i586-poky-linux-libtool: link: See the libtool documentation for more information.
| i586-poky-linux-libtool: link: Fatal configuration error.
as the hardcoded test result breaks libtool internal assumptions about linking on
Linux.
Old 1.6.8p12 is quite old, even w/o source available now (only cached in pokylinux).
[Patches-Poky]
KEEP _autofoo.patch_: deal with bunch of autotool warnings/errors, more about
underquotes. In 1.7.2p7, those error conditions disappear.
But the underquotes fix is still valid to remove a bunch
of autoconf warnings
REMOVE _autofoo1.patch_: copy of autofoo.patch. no one uses it. REMOVE.
KEEP _noexec-link.patch_: use "-module" to avoid libtool complaining about not
starting with "lib". Add comment to the patch file.
DISABLE _nostrip.patch_: it's wrong to strip target binaries on build host. Borrow
from OE to use 'sed' instead of patch, but keept it for
remind to upstream
REMOVE _nonrootinstall.patch_: handle mode/owner population at install. It's not
requried at all since there's pkg_postinst to handle
it at right stage
Other misc stuff:
- update metadata information
- no need to use ${S} in 1.7.2p7
- fix idention
- use sed to disable strip
- changes from OE:
* disable PAM as our pam doesn't include necessary config files and plugs
* no need to special sudo_cv_uid_t_len tweak in the recipe. It's already there
in poky site files for a long time
* move sudo_cv_func_unsetenv_void to meta/site/common. Stdlib.h is common across
architectures in our libc headers, which conforms to POSIX standard
(borrow from openembedded 1.7.2p4)
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
Richard Purdie
Lukas Bulwahn
Kevin Tian
Richard Purdie
Richard Purdie
Ross Burton