Bug#441975: libc6: Still present !!!
Gallien Matthieu a écrit :
> Le Monday 14 January 2008 16:49:01 Aurelien Jarno, vous avez écrit :
>> Matthieu Gallien a écrit :
>>> Le lundi 14 janvier 2008, Aurelien Jarno a écrit :
>>>> tag 441975 + unreproducible
>>>> thanks
>>>>
>>>> Gallien Matthieu a écrit :
>>>>> Package: libc6
>>>>> Version: 2.7-6
>>>>> Followup-For: Bug #441975
>>>>>
>>>>> I do not know what to say.
>>>>> The bug is still here.
>>>>> Just got :
>>>>> Préconfiguration des paquets...
>>>>> (Lecture de la base de données... 320682 fichiers et répertoires déjà
>>>>> installés.)
>>>>> Préparation du remplacement de login 1:4.1.0-1 (en utilisant
>>>>> .../login_1%3a4.1.0-2_i386.deb) ...
>>>>> Dépaquetage de la mise à jour de login ...
>>>>> Paramétrage de login (1:4.1.0-2) ...
>>>>> (Lecture de la base de données... 320682 fichiers et répertoires déjà
>>>>> installés.)
>>>>> Préparation du remplacement de ncurses-bin 5.6+20071215-1 (en utilisant
>>>>> .../ncurses-bin_5.6+20080105-1_i386.deb) ...
>>>>> Dépaquetage de la mise à jour de ncurses-bin ...
>>>>> Paramétrage de ncurses-bin (5.6+20080105-1) ...
>>>>> (Lecture de la base de données... 320682 fichiers et répertoires déjà
>>>>> installés.)
>>>>> Préparation du remplacement de libc6-dev 2.7-5 (en utilisant
>>>>> .../libc6-dev_2.7-6_i386.deb) ...
>>>>> Dépaquetage de la mise à jour de libc6-dev ...
>>>>> Préparation du remplacement de libc6-dbg 2.7-5 (en utilisant
>>>>> .../libc6-dbg_2.7-6_i386.deb) ...
>>>>> Dépaquetage de la mise à jour de libc6-dbg ...
>>>>> Préparation du remplacement de libc6 2.7-5 (en utilisant
>>>>> .../archives/libc6_2.7-6_i386.deb) ...
>>>>> Dépaquetage de la mise à jour de libc6 ...
>>>>> Paramétrage de libc6 (2.7-6) ...
>>>>> dpkg : erreur de traitement de libc6 (--configure) :
>>>>> sous-processus post-installation script tué par le signal (Erreur de
>>>>> segmentation)
>>>>> Des erreurs ont été rencontrées pendant l'exécution :
>>>>> libc6
>>>>> E: Sub-process /usr/bin/dpkg returned an error code (1)
>>>>> wahiba# perl perl -e "use Qt"
>>>>> Can't open perl script "perl": Aucun fichier ou répertoire de ce type
>>>>> wahiba# perl -e "use Qt"
>>>>> zsh: segmentation fault perl -e "use Qt"
>>>>> wahiba# rm /etc/ld.so.nohwcap
>>>>> wahiba# perl -e "use Qt"
>>>>>
>>>>> The last call has no problem to complete without segfault.
>>>>>
>>>>> I have no the technical skills and the time solve this, but can you
>>>>> please do it.
>>>> Unfortunately nobody can reproduce the bug, so nobody can fix it...
>>> So, you say that bug 453480 does not exist ?
>> Maybe it exists, but not on the machines (4 in total) I have tested. I
>> know some other persons also have tried to reproduce it and failed.
>>
>>> In this case I will stop do any report on that bug and live with it as
>>> there is a workaround that works for me. But this annoying ...
>>> In the other case, I believe that the bug was reproduced and not fixed.
>> It seems a few other people have encountered it, but that doesn't make
>> it fixable. We have to reproduce it to be able to debug and fix it.
>
> So what you say is that you cannot reproduce but others can. It seems for me
> different from unreproducible that give the message that it is not a real bug
> for other users ...
Not "others", but only a few people, and unfortunately they don't have
the skill to debug it.
> This is my last words about this bug :
> Is it not possible to check if the specific libc is installed when upgrading
> and thus does not create the infamous file that cause the bug for me ?
This file is created *because* you have libc6-i686. If it is not
created, the critical part of the glibc unpacking could crash, rendering
your machine unusable.
> I do not know the internal of the library loader of linux and so cannot do a
> proper fix by miself.
>
> Finally I would say that I was friendly and offered help to find a solution. I
> could test your fix, try specific solutions or workaround. Even I could have
Yes I agree, and we even ask you to do some basic tests, and you answer
us. Unfortunately, we've reach a point when it is not possible to debug
blindly and we need to be able to reproduce the problem.
> tried to offer access to my hardware, if it is needed.
Is it really possible? People usually refuse to give an access to their
machine, so I don't even tried to ask that.
--
.''`. Aurelien Jarno | GPG: 1024D/F1BCDB73
: :' : Debian developer | Electrical Engineer
`. `' aurel32@debian.org | aurelien@aurel32.net
`- people.debian.org/~aurel32 | www.aurel32.net
Reply to: