[blfs-dev] glibc errors in blfs packages on i686

Fernando de Oliveira famobr at yahoo.com.br
Sun Sep 1 04:59:38 PDT 2013


Em 31-08-2013 21:16, Bruce Dubbs escreveu:
> Ken Moffat wrote:
>> On Fri, Aug 30, 2013 at 09:20:28AM -0500, Bruce Dubbs wrote:
>>>
>>> OK, I just rebuilt LFS in my sandbox with the new packages.  It was
>>> extremely clean.  I had no FAILs at all and The only Errors were from glibc:
>>>
>>> make[3]: *** [/sources/glibc-build/posix/tst-getaddrinfo4.out] Error 1
>>> make[3]: [/sources/glibc-build/posix/annexc.out] Error 1 (ignored)
>>> make[3]: [/sources/glibc-build/conform/run-conformtest.out] Error 1
>>> (ignored)
>>>
>>> OTOH, I understand the concerns.  I'll revert what I did in my sandbox
>>> and rebuild and from that do -rc2.  I will still take the advice in
>>> Chapter 3:
>>>
>>> "The Linux kernel is updated relatively often, many times due to
>>> discoveries of security vulnerabilities. The latest available 3.10.x
>>> kernel version should be used"
>>>
>>>     -- Bruce
>>
>>   In -rc2 I had my normal fails in binutils and parts of gcc (that
>> was with my usual "set CFLAGS, CXXFLAGS and sweep the static libs
>> under the carpet, also build using all CPUs" variant.  My glibc
>> results looked like yours.
>>
>>   Anyway, I've now completed my normal desktop on x86_64 with intel
>> graphics.  In case there is any doubt, I am standing by everything I
>> tagged for 7.4 in -rc1, with the exception of two packages (elfutils
>> and the ati video driver) that I only build on desktops with an ATI
>> video chip.  From here it is looking very good.
>>
>>   Thanks for your work in identifying the i686 problem.
> 
> That was Fernando.  I think him for spotting the problem.

Unfortunately, it happened to me, being the first one having a system
affected. After two more lfs 7.4rc1 built (all three using jhalfs) and
the same error message involving glibc and java (several binaries
tested, including a proprietary one), I was *suspicious* about glibc,
but not wanting to risk being wrong about it and upset Bruce and Matt. I
narrowed down to two possible sources of the problem: jhalfs or glibc.
If jhalfs, it could be my mistake. But failed to find the discussion in
Arch, because was always wondering "why only java is affected, nothing
else" and searched the web always including openjdk. Then, Bruce made
the right search and easily found the discussion and work around. It is
described there that other packages are also affected (although one
reported to be affected, glxgears, worked for me). I will try to be
smarter, next time (hopefully, no next time will happen), so that we do
not have to delay so much postpone a release.

> 
> With everyone's help, we can get out a good release.  I'm especially 
> happy with everyone's contributions to BLFS.  I think it will be a major 
> accomplishment if we can get out the first stable BLFS release since 
> August 2008.
> 
>    -- Bruce

Thanks, Bruce. I am really excited with only the idea of having a new
stable BLFS, more yet in being part of it.


-- 
[]s,
Fernando



More information about the blfs-dev mailing list