Resent without the "SPAM" header, sorry for the first one.

On 04/04/2018 09:27, Pierre Labastie wrote:
> On 04/04/2018 01:48, Alain Toussaint wrote:
>> Le mercredi 04 avril 2018 à 00:44 +0100, Ken Moffat a écrit :
>>> On Tue, Apr 03, 2018 at 05:46:49PM -0400, Alain Toussaint wrote:
>>>> Hello,
>>>>
>>>> I successfully built llvm by hand after trying to do so with ALFS-svn 
>>>> checkout on the Sunday 1st
>>>> of
>>>> April with both LFS and BLFS checked out on the same date. The ALFS logs 
>>>> for llvm show this
>>>> issue:
>>>>
>>>> In file included from 
>>>> /sources/llvm/llvm-6.0.0.src/build/include/llvm/Support/DataTypes.h:33:0,
>>>>                  from 
>>>> /sources/llvm/llvm-6.0.0.src/include/llvm/Support/SwapByteOrder.h:19,
>>>>                  from 
>>>> /sources/llvm/llvm-6.0.0.src/include/llvm/Support/MathExtras.h:18,
>>>>                  from 
>>>> /sources/llvm/llvm-6.0.0.src/include/llvm/ADT/APInt.h:20,
>>>>                  from 
>>>> /sources/llvm/llvm-6.0.0.src/include/llvm/ADT/APFloat.h:20,
>>>>                  from 
>>>> /sources/llvm/llvm-6.0.0.src/lib/Support/APFloat.cpp:15:
>>>> /usr/include/c++/7.3.0/cmath:45:15: erreur fatale: math.h : Aucun fichier 
>>>> ou dossier de ce type
>>>>  #include_next <math.h>
>>>>
>>>
>>> That sounds suspiciously like your mesa problem in February.  After
>>> someone hit a similar problem on support this week, I found the past
>>> posts.  I think you have recompiled gcc in BLFS before you got to
>>> this problem ?
>>
>> Exactly the same. Matter of fact, it did occur right now with poppler and 
>> mesa. I compile them
>> manually (not a major hindrance) and I'm able to get past.
>>
>> both hand compilation and ALFS are done as a user (alain) with gid=users but 
>> then I suspect ALFS
>> modifies some environment variables.
>>
> 
> I really do not see which envar could be set, which could explain that. All
> the envars from jhalfs (the blfs tools at least) are set in envars.conf, and
> in scriptlets. If you have a recent enough version of jhalfs, the envar names
> are all prefixed with "JH_", in order to prevent conflict with system envars.
> 
> But I'd like to help debug this problem. Could you please, in a shell where
> building manually works, run:
> $ source <envars.conf directory>/envars.conf
> 
> then build manually again, and report if it still works or not?
> 
> Pierre
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Reply via email to