[oe] [PATCH][meta-oe] crash: fix do_compile relocation error

[Changqing Li]

On 07/03/2018 12:42 AM, Khem Raj wrote:
> On Mon, Jul 2, 2018 at 12:08 AM Changqing Li <changqing.li at windriver.com> wrote:
>> On 07/02/2018 02:09 PM, Khem Raj wrote:
>>> On Sun, Jul 1, 2018 at 7:31 PM Changqing Li <changqing.li at windriver.com> wrote:
>>>> error info:
>>>>
>>>> id: relocation error: /lib/x86_64-linux-gnu/libnss_nis.so.2:
>>>> symbol _nsl_default_nss version GLIBC_PRIVATE not defined in
>>>> file libnsl.so.1 with link time reference
>>>>
>>>> Recent glibc change removed libnss-nis module from glibc and
>>>> a new recipe libnss-nis.bb was added. After this change,
>>>> we need to include native libnss_nis in crash. otherwise,
>>>> with "nis" in /etc/nssswitch.conf, we will met above error．
>>>>
>>>> Signed-off-by: Changqing Li <changqing.li at windriver.com>
>>>> ---
>>>>    meta-oe/recipes-kernel/crash/crash_7.2.0.bb | 1 +
>>>>    1 file changed, 1 insertion(+)
>>>>
>>>> diff --git a/meta-oe/recipes-kernel/crash/crash_7.2.0.bb b/meta-oe/recipes-kernel/crash/crash_7.2.0.bb
>>>> index 440a69d..09fe50d 100644
>>>> --- a/meta-oe/recipes-kernel/crash/crash_7.2.0.bb
>>>> +++ b/meta-oe/recipes-kernel/crash/crash_7.2.0.bb
>>>> @@ -11,6 +11,7 @@ LICENSE = "GPLv3"
>>>>    LIC_FILES_CHKSUM = "file://COPYING3;md5=d32239bcb673463ab874e80d47fae504"
>>>>
>>>>    DEPENDS = "zlib readline coreutils-native"
>>>> +DEPENDS_append_libc-glibc = " libnss-nis-native"
>>> A native dependency which is specific to a target libc seems not
>>> right. It should depend on
>>> build host libc which in this case is always glibc so please drop the override.
>> Hi,  according  to below content of recipe libnss-nis,  only for glibc,
>> we will DEPNDS
>> libnss-nis-native, so I add this override.
>>
>> # We will skip parsing this packagegeoup for non-glibc systems
>> #
>> python __anonymous () {
>>       if d.getVar('TCLIBC') != "glibc":
>>           raise bb.parse.SkipRecipe("incompatible with %s C library" %
>>                                      d.getVar('TCLIBC'))
> may be this can be made to exclude non target recipes.
>

I don't quite get your mean, can you explain more? Thanks.