[arch-dev-public] WARNING - glibc-2.13-2

Allan McRae allan at archlinux.org
Sat Feb 5 20:57:12 EST 2011

On 06/02/11 11:02, Thomas Bächler wrote:
> Am 05.02.2011 12:24, schrieb Allan McRae:
>> On 05/02/11 20:20, Ionuț Bîru wrote:
>>> On 02/05/2011 12:03 PM, Allan McRae wrote:
>>>> New upstream snapshot. Some useful patches have been pulled to the 2.13
>>>> branch.
>>>> The only new testsuite failures are from new additions to the testsuite
>>>> that do not like being built with -Wl,--as-needed.
>>>> Signoff both,
>>>> Allan
>>> no signoff from me
>>> xchat
>>> Inconsistency detected by ld.so: dl-deps.c: 622: _dl_map_object_deps:
>>> Assertion `nlist>  1' failed!
>>> works fine once i downgrade to core
>> Yay for backports that break working stuff!
>> http://sourceware.org/bugzilla/show_bug.cgi?id=12454
>> Will be looking out for the fix to be approved.
>> Allan
> The package shouldn't stay in testing that way - this just came from IRC:

I'm deciding what is best to do here.   I could revert the commit that 
causes this issue but it is actually the most important backport onto 
the glibc-2.13 branch so far...  There is also a patch on glibc bugzilla 
that looks like a fix.  I would like to see the upstream response before 
pulling it, but might pull it anyway and push a new package to [testing] 
in the knowledge that I might have to revert that patch for the approved 
fix later.  Expect some kind of fix in the next day once I have done 
some more testing.

> 'hello, i am getting "Inconsistency detected by ld.so: dl-deps.c: 622:
> _dl_map_object_deps: Assertion 'nlist>  1' failed!" when starting X.org.
> I am using testing. ideas on how to make x.org work ?'

What ever happened to [testing] users reading this mailing list and 
being competent enough to downgrade if they are affected by a bug in a 
package?  Maybe the big WARNING I added to the subject line will help! :P

> If I understand this right, it'll break everything :)

It probably won't break everything...   It is a weird bug that affects 
some programs for some people but not others.  It was unfortunate that 
my system seems completely unaffected, including not getting the glibc 
test suite failure that could have detected this.


More information about the arch-dev-public mailing list