Age | Commit message (Collapse) | Author | Files |
|
* install MakeMaker files separately for native and target system
* PERL_INC, PERL_LIB and PERL_ARCHLIB changes through environment
* for more see http://lists.linuxtogo.org/pipermail/openembedded-devel/2008-November/006815.html
|
|
Conflicts:
packages/busybox/busybox_1.11.3.bb
packages/dsplink/codec-engine_2.10.bb
packages/gdb/gdb_6.3.bb
packages/gdb/gdb_6.4.bb
packages/gdb/gdb_6.6.bb
packages/gdb/gdb_6.8.bb
packages/glibc/glibc_2.2.5.bb
packages/glibc/glibc_2.3.2+cvs20040726.bb
packages/glibc/glibc_2.3.2.bb
packages/glibc/glibc_2.3.3+cvs20041128.bb
packages/glibc/glibc_2.3.3+cvs20050221.bb
packages/glibc/glibc_2.3.3+cvs20050420.bb
packages/glibc/glibc_2.3.3.bb
packages/glibc/glibc_2.3.5+cvs20050627.bb
packages/glibc/glibc_2.3.6.bb
packages/glibc/glibc_2.4.bb
packages/glibc/glibc_2.5.bb
packages/glibc/glibc_2.6.1.bb
packages/glibc/glibc_2.7.bb
packages/kexecboot/linux-kexecboot_2.6.26.bb
packages/libsdl/libsdl-x11_1.2.11.bb
packages/linux/linux-rp_2.6.26.bb
packages/mozilla/fennec_hg.bb
packages/python/python_2.5.2.bb
packages/tasks/task-xfce-base.bb
packages/xorg-driver/xf86-video-omapfb_git.bb
packages/xorg-xserver/xserver-kdrive_1.4.0.90.bb
|
|
|
|
This is a partial revert of aa03004ace2fabb135a6208ef8c2d2b312aa7b0b. The
approach that was taken leads to disagreement and the change was pushed
too early. DISTRO_PR is going to say and will be used inside the package
creation. There will be more disucssion on the list about where to continue
from here.
|
|
Allow a distribution to globally bump the PR of every package
this is useful when there was a change in the toolchain and
every package is going to be different.
find packages/ \( -name '*.bb' -or -name '*.inc' \) -exec sed -i s/"^PR="/"FILE_PR ="/ {} \;
find packages/ \( -name '*.bb' -or -name '*.inc' \) -exec sed -i s/"^PR ="/"FILE_PR ="/ {} \;
Acked-By: Koen Kooi <koen@openembedded.org>
|
|
|
|
STAGING_DIR_HOST for the upcoming MULTIMACH_TARGET_SYS changes
* see http://lists.linuxtogo.org/pipermail/openembedded-devel/2008-July/005819.html
|
|
Bump PR
|
|
host env
* CAVEAT: perl/5.8.8 has to be hard coded, needs to be updated with each new perl version
|
|
host env
* CAVEAT: perl/5.8.8 has to be hard coded, needs to be updated with each new perl version
|
|
|
|
|
|
* bugs.openembedded.org has frequent DNS issues, .net is better maintained
both should usually point to the same server and bugzilla installation
|
|
60d4d904e10f8212b1c2242377b31ccc2e29e374 | patch -R -p0 --remove-empty-files
disapprove for the poor and "mtn suspend"-les people
|
|
|
|
|
|
|
|
|
|
|
|
Update the sed magic to catch the files that were being missed. Closes #2575.
|
|
it in order to get the definition for getpagesize which has been definied in
unistd.h, which it already includes, since glibc 2.1. Some versions of linux
libc headers removed the asm/page.h resulting in failures here for some
people. Closes #2546.
|
|
|
|
set to point at the current perl build to find the shared perl library,
however in the OE case we actually want the staged native version not those
that belong to the build for the target. The miniperl binary actually has an
rpath pointing at the staging dir so no LD_LIBRARY_PATH is required. It
seems gentoo has started letting LD_LIBRARY_PATH override rpath which is
breaking the perl build. Should solve part of #2483 and other irc reports of
problems.
|
|
makefile within the perl build apply the patches. This was a left over from
the previous perl version and makes modifying these patches a lot more
difficult than it really needs to be.
|
|
seperating space in some cases. Report as part of #2483.
|
|
switches for some hacked gcc version, which are not recognized with official
gcc's.
* See #1980.
|
|
requested in bug #1980. Closes #1980.
|
|
was already included being included in perl-native, but not in perl itself.
|
|
* DEPENDS on perl-native isn't needed since cpan.bbclass adds this.
* DEPENDS on expat-native isn't needed since libxml-parser-perl adds this.
* RDEPENDS on perl-native was only here to prevent an RDEPENDS on perl. The
RDEPENDS was being added by cpan.bbclass, but that has now been fixed to
not add it for native packages. This needed to be removed so that perl
isn't built when you only want libxml-parser-perl-native.
|
|
* Stage the headers from perl
* Modify cpan.bbclass to use the appropriate headers.
* Remove prototypes as well as disabling functions that uclibc doesn't
support.
* Bump PR on libxml-parser-perl to force it to be rebuilt.
Previously it was always using the headers from native-perl, since the
headers were the same for native and target... but that's only true in the
case of glibc, for uclibc the headers are different.
|
|
|
|
to take into account the new staging location of perl. This will also force
a rebuild of libxml-parser-perl which should now work.
|
|
properly terminating a quote. This is in theory the fix for the /bin/sh ->
/bin/dash problems with perl, however since I don't have any issues here with
dash (ubuntu 6.10) I can't confirm this. The patch is an obvious fix though
so I'm adding it.
|
|
environment and use that instead of the configured location of PERL_INC.
Without this PERL_INC for non-native recipes ends up pointing at where the
headers will be located on the target. In theory we could override this in
cpan class when calling build.PL but for some packages, such as
libxml-parser-perl, that only fixes the top-level makefile and is not
propagated to the sub-makefiles. This change results in MakeMaker always
picking up the correct staged location of the include files without
effecting where they expect to be found on the target.
|
|
perl-native to remove the various directories and to remove the host
specific path components. This is to make it easier to refer to the staged
perl location from other places, such as when trying to tell things where
the perl include files are located.
|
|
else happens to want to use them.
|
|
been effected by the revent CFLAGS changes to perl 5.8.8.
|
|
the CFLAGS from the perl configuration with the OE CFLAGS we prepend the OE
CFLAGS. This is needed to ensure that compiler defines (such as _GNU_SOURCE)
are correctly definied when building perl modules. This primarily effects
modules that link against the core perl library, such as libxml-parser-perl.
|
|
|
|
with the new version of perl.
|
|
Some distributions have already started to use this so it's probably about
time to promote it to the default.
|
|
cause issues with upgrading from earlier 5.8.8 versions (perl-lib will
remain instead of upgrading to libperl5) but it'll actually make upgrades
from 5.8.7 work. I can't see how to fix the 5.8.8 upgrading easily. The
renaming was disabled due to a mistake on my part.
|
|
perl builds on some hosts.
|
|
perl-native configure script was added -I/usr/include/gdbm to the ccflags if
you host had a /usr/include/gdbm/ndbm.h header, even though gdbm support is
disabled. This patch stops the configure script adding that even if it finds
the header there.
|
|
cpan-base.bbclass, and then update the cpan modules to no longer manually
specify this. Also remove references to the new removed .packlist files.
|
|
from being installed. Since we use perl-native to do the cpan module
installs we need to changed perl-native as well to stop cpan modules from
getting .packlist files. This is a new patch since just the .packlist
related part has been extracted from the larger patch.
|
|
from the configure task in perl to the stage task in perl-native where it
belongs - perl shouldn't be messing with the staged perl-native.
|
|
cpan.bbclass changes.
|
|
the PACKAGES_DYNAMIC declaration for the perl modules. This declaration
appears to be global though - so if either of the older perl versions exist
it actually works - bitbake picks up the definition from the older recipes
even though it isn't building them. This shows up as "No providers of
runtime build target <x>" errors where x is a perl module. Thanks to Koen
for figuring this one out.
|
|
This is done by allowing gcc to search for errno.h by itself instead
of manually searching for. The manual search was looking in
STAGING_INCDIR and that's not where the external toolchains headers
are. This whole test is really for handle other compilers and
operating systems, so the simple make gcc do itself should be fine
for us. Closes #1700.
|