On Wed, Dec 16, 2015 at 7:50 AM, Aníbal Limón <anibal.li...@linux.intel.com> wrote: > Hi Matt, > > I tried to do again without your patches and the problem appears, seems > to be an older problem.
Thanks, I suspected as much and was going to try that myself. I was able to work around the issue manually in a devshell by using apt-get -f to have it "fix" the dependencies it thinks were broken, which from what I've seen so far, stems from trying to install 32-bit packages that would overwrite files installed by their 64-bit counterparts. I'm just not sure that's the right solution to the problem. -Matt > > Regards, > alimon > > On 12/16/2015 07:31 AM, Matt Madison wrote: >> On Tue, Dec 15, 2015 at 2:23 PM, Aníbal Limón >> <anibal.li...@linux.intel.com> wrote: >>> Hi Matt, >>> >>> Trying to build core-image-sato with qemux86-64 and multilib enabled an >>> errors appear (see attached log), next the config. >> >> OK, yep, I can reproduce this. I'll investigate further. >> >> Thanks, >> -Matt >> >> >>> >>> MACHINE ??= "qemux86-64" >>> >>> IMAGE_INSTALL_append = " lib32-connman" >>> require conf/multilib.conf >>> MULTILIBS = "multilib:lib32" >>> DEFAULTTUNE_virtclass-multilib-lib32 = "x86" >>> PACKAGE_CLASSES ?= "package_deb" >>> EXTRA_IMAGE_FEATURES = "debug-tweaks package-management" >>> >>> Kind regards, >>> alimon >>> >>> >>> On 12/15/2015 01:28 PM, Matt Madison wrote: >>>> On Tue, Dec 15, 2015 at 9:29 AM, Aníbal Limón >>>> <anibal.li...@linux.intel.com> wrote: >>>>> Hi Matt, >>>>> >>>>> I'm starting to look at your patches, in what arches/combinations you >>>>> test the patches? >>>> >>>> I've been working on a BSP layer for the jetson-tx1, which is >>>> aarch64/armv7a-hf, so that's what I've been testing with. The BSP >>>> defs are in https://github.com/madisongh/meta-tegra. >>>> >>>> Thanks, >>>> -Matt >>>> >>>> >>>>> >>>>> Kind regards, >>>>> alimon >>>>> >>>>> >>>>> On 12/06/2015 11:25 AM, Matt Madison wrote: >>>>>> I ran into sevearl issues while trying to build an ARM multilib rootfs >>>>>> using Debian packaging. After several go-rounds, it looked like the >>>>>> cleanest solution was to tweak how DPKG_ARCH gets constructed and >>>>>> to have the DpkgPM class in oe/package_manager.py use that variable >>>>>> to locate multilib variants (similar to RpmPM). I also took the >>>>>> liberty of expanding the Debian architecture mappings so the names >>>>>> align better with what's documented on the Debian wiki, for those >>>>>> cases where a direct mapping is possible. >>>>>> >>>>>> Matt Madison (2): >>>>>> package_deb.bbclass, cross-canadian.bbclass: DPKG_ARCH mapping >>>>>> function >>>>>> package_manager.py: fixes for multilib deb packaging builds >>>>>> >>>>>> meta/classes/cross-canadian.bbclass | 2 +- >>>>>> meta/classes/package_deb.bbclass | 35 >>>>>> +++++++++++++++++++++++++---------- >>>>>> meta/lib/oe/package_manager.py | 17 +++++++++++------ >>>>>> 3 files changed, 37 insertions(+), 17 deletions(-) >>>>>> -- _______________________________________________ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core