[lfs-dev] Possible problem with current glibc (LFS 7.2 cant recompile LFS 7.2)

Jasmine Iwanek jasmine at iwanek.co.uk
Sat Aug 25 16:39:42 PDT 2012


On 2012-08-26 00:26, William Harrington wrote:
> On Aug 25, 2012, at 17:43 PM, Bruce Dubbs wrote:
>
>> Jasmine Iwanek wrote:
>>> On 2012-08-25 22:35, Bruce Dubbs wrote:
>>>>
>>>> OK, that's pretty early in the build.  Does /usr/include/rpc/rpc.h
>>>> exist
>>>> on the host?
>>>>
>>>
>>> Not unless I add it. (Which also fixes it by the way), bit out of
>>> scope
>>> however, after all there's nothing installing it in 7.2 dev, and
>>> while
>>> it might be mentioned in BLFS, That's outside of the scope of
>>> bootstrapping 7.2 with just 7.2.
>>>
>>>> If you add the following to the start of the chapter 5 
>>>> instructions,
>>>> does it build?
>>>>
>>>> sed -i 's|<rpc/types.h>|"rpc/types.h"|' sunrpc/rpc_clntout.c
>>>>
>>>
>>> I'd given an update in my other message, but as you were replying
>>> as I
>>> was, yes, works fine with that sed.
>>
>> OK, we can add that to the current book before we go to a final 7.2.
>>
>>    -- Bruce
>>
>
> Hello Bruce,
>
> I ran across this in our CLFS multilib builds when during 32 bit
> eglibcbuilds, it is cross compiling using a 64 bit host and thus it
> puts glibc/eglibc in a bad state and thus using internal rpc headers
> is required when they are not available on the host. That's why I
> suggested that earlier in the thread.
>
> You will not encounter this when not cross compiling at all.
>
> Sincerely,
>
> William Harrington

Interestingly My host is 32bit, the LFS target is 32bit,
I would suspect this is triggered in 7.2 by the faked crossbuild.

--
Jasmine Iwanek




More information about the lfs-dev mailing list