Re: Debian 9.2 amd64 Xfce lock screen -> 100 %CPU by lightdm-gt+

[davidson]

On Wed, 11 Oct 2017, David Christensen wrote:


debian-user:

I have a Dell Inspiron 6400 laptop with:

2017-10-11 20:06:21 dpchrist@tinkywinky ~
$ cat /etc/debian_version
9.2

2017-10-11 20:07:34 dpchrist@tinkywinky ~
$ uname -a
Linux tinkywinky 4.9.0-4-amd64 #1 SMP Debian 4.9.51-1 (2017-09-28) x86_64 
GNU/Linux



Recently, I've noticed that if I lock the screen and walk away, when I return 
the CPU fan is on high speed.  When I unlock the screen, the CPU Graph 
indicates that one of my two CPU cores was pegged at 100%, but dropped off 
once I logged in.



If I lock the screen and SSH in from another machine, 'top' says:

top - 20:02:15 up 12:54,  1 user,  load average: 0.28, 0.42, 0.73
Tasks: 217 total,   3 running, 214 sleeping,   0 stopped,   0 zombie
%Cpu(s): 45.6 us,  6.3 sy,  0.0 ni, 48.2 id,  0.0 wa,  0.0 hi,  0.0 si, 0.0 
st

KiB Mem :  2043196 total,   820100 free,   661892 used,   561204 buff/cache
KiB Swap:  1952764 total,  1823528 free,   129236 used.  1145664 avail Mem

 PID USER  PR  NIVIRTRESSHR S  %CPU %MEM TIME+ COMMAND
7473 lightdm   20   0  587348  51332  25584 R  94.6  2.5   0:14.52 
lightdm-gt+

7462 root  20   0  259828  19656  13144 S  10.6  1.0   0:01.66 Xorg
2463 root  20   0  349180  44600   6068 S   0.4  2.2   3:20.36 mongod
 514 mongodb   20   0  322476  11664   5328 R   0.2  0.6   2:59.51 mongod
 638 root  20   0  384964  24920  15168 S   0.2  1.2   6:34.34 Xorg
 886 root  20   0 780  10052   7872 S   0.2  0.5   0:26.93 jsvc
7455 dpchrist  20   0   43272   3344   2768 R   0.2  0.2   0:00.06 top
 993 root  20   0 4413496 289804  10428 S   0.1 14.2   1:36.59 java
1113 dpchrist  20   0  191548  17040  12000 S   0.1  0.8   1:01.25 xfwm4
...


So, lightdm-gt+ is the culprit.


AFAICT, that command name is truncated. In interactive top, you can
scroll with right-arrow, to view the rest of the name. With that
information, your search results might improve.

Could be lightdm-gtk-greeter? Or something else, I guess.

If/when I run the laptop on the battery, this will be a serious problem 
(especially given a weak old battery).



I don't see this issue in the 'lightdm' bug list:

https://bugs.debian.org/cgi-bin/pkgreport.cgi?dist=unstable;package=lightdm


If the culprit is lightdm-gtk-greeter, it has its own eponymous package.

Hope this helps (a little).

Re: RE

[Jape Person]

On 10/11/2017 08:48 PM, Jimmy Johnson wrote:
> On 10/11/2017 10:14 AM, Jape Person wrote:
>> On 10/11/2017 04:20 AM, debian-user@lists.debian.org wrote:
>>> Check
>>> Re: hplip and use of the "driver plugin"
>>>
>>
>> I'm seeing something from debian-user@lists.debian.org that I've
>> never seen before. The most recent three days I have received a
>> single e-mail as quoted above.
>>
>> Is this a normal function of the list? If so, what purpose does
>> it serve? The thread referenced died out in December of last year.
>>
>> I just started getting back into the swing of things after a
>> prolonged medical issue, but didn't see any new messages in the
>> thread when I came back to check my mail. Is my attention being
>> drawn to the thread for some specific reason?
>>
>> Thanks to anyone who can provide a clue!
>>
>> JP
> 
> It looks like they are trying to get an update to: 
> https://lists.debian.org/debian-user/2016/11/msg00769.html
> Maybe a bug was filed on your behalf, anyways it looks like someone is 
> interested in some kind of closure.  Just so you know you're not the 
> only one, I had the same thing going on and gave them some feed back on 
> a i965 video driver in stretch while in testing and I closed the bug 
> cause it was working, I hope this helps. Also it shows that Debian is 
> listening to you.
> 
Thanks, Jimmy.

I did already look through the thread, but I think the last post
to it was 7th December, 2016.

I also found bug #451502 from 2007.

hplip: may download non-free plugins

It got considered for a couple of years and was finally answered
with a wontfix.

I understand the upstream hplip people and the package
maintainers for Debian and their decision to keep hplip in main.

My response to HP (the company, not any of the folks involved in
trying to make all of the new HP devices work in Debian) is --

wontbuy.



Instead, I got a Brother MFC 9340CDW and am able to use all of
its features (including scanning via a weird workaround) without
installing anything from any source other than Debian main.

If Brother ever fixes the firmware so that it reports the
printer's actual resolution, I'll even be able to use the IPP /
driverless printing functionality and dispense with the PPD for
the 9320CD that I currently have to use in CUPS to get proper
print output.

Right now the latest firmware available is 1.09, and it causes
driverless printing to result in squashed output.

But at least this printer doesn't require me to stick a
proprietary blob on my system drive to get scanning to work, the
way most of the current HP non-inkjet MFCs do.

Thanks for your response!

JP

PS: My sincere apologies to Jimmy Johnson. Not sure how it
happened. I accidentally replied, at first, directly to Jimmy
instead of to the list. The old eyes and hands just don't
coordinate well any more.

Debian 9.2 amd64 Xfce lock screen -> 100 %CPU by lightdm-gt+

[David Christensen]

debian-user:

I have a Dell Inspiron 6400 laptop with:

2017-10-11 20:06:21 dpchrist@tinkywinky ~
$ cat /etc/debian_version
9.2

2017-10-11 20:07:34 dpchrist@tinkywinky ~
$ uname -a
Linux tinkywinky 4.9.0-4-amd64 #1 SMP Debian 4.9.51-1 (2017-09-28) 
x86_64 GNU/Linux



Recently, I've noticed that if I lock the screen and walk away, when I 
return the CPU fan is on high speed.  When I unlock the screen, the CPU 
Graph indicates that one of my two CPU cores was pegged at 100%, but 
dropped off once I logged in.



If I lock the screen and SSH in from another machine, 'top' says:

top - 20:02:15 up 12:54,  1 user,  load average: 0.28, 0.42, 0.73
Tasks: 217 total,   3 running, 214 sleeping,   0 stopped,   0 zombie
%Cpu(s): 45.6 us,  6.3 sy,  0.0 ni, 48.2 id,  0.0 wa,  0.0 hi,  0.0 si, 
0.0 st

KiB Mem :  2043196 total,   820100 free,   661892 used,   561204 buff/cache
KiB Swap:  1952764 total,  1823528 free,   129236 used.  1145664 avail Mem

  PID USER  PR  NIVIRTRESSHR S  %CPU %MEM TIME+ COMMAND
 7473 lightdm   20   0  587348  51332  25584 R  94.6  2.5   0:14.52 
lightdm-gt+

 7462 root  20   0  259828  19656  13144 S  10.6  1.0   0:01.66 Xorg
 2463 root  20   0  349180  44600   6068 S   0.4  2.2   3:20.36 mongod
  514 mongodb   20   0  322476  11664   5328 R   0.2  0.6   2:59.51 mongod
  638 root  20   0  384964  24920  15168 S   0.2  1.2   6:34.34 Xorg
  886 root  20   0 780  10052   7872 S   0.2  0.5   0:26.93 jsvc
 7455 dpchrist  20   0   43272   3344   2768 R   0.2  0.2   0:00.06 top
  993 root  20   0 4413496 289804  10428 S   0.1 14.2   1:36.59 java
 1113 dpchrist  20   0  191548  17040  12000 S   0.1  0.8   1:01.25 xfwm4
...


So, lightdm-gt+ is the culprit.


If/when I run the laptop on the battery, this will be a serious problem 
(especially given a weak old battery).



I don't see this issue in the 'lightdm' bug list:

https://bugs.debian.org/cgi-bin/pkgreport.cgi?dist=unstable;package=lightdm


STFW yielded no hits.


Has anyone else seen this?


Any suggestions for debugging or fixing it?


Should I file a bug report against 'lightdm'?


David

Re: "spelling check" pour le français

[Yuwen Dai]

Merci Bernard, mais mon question est pourquoi "allez" est un invalide mot.
Vous n'avez pas le même problème?

slt
Yuwen

Le 12 octobre 2017 à 10:35,  a écrit :

>
>
> --
>
> *De: *"Yuwen Dai" 
> *À: *debian-user-french@lists.debian.org
> *Envoyé: *Jeudi 12 Octobre 2017 04:24:18
> *Objet: *"spelling check" pour le français
>
>
> Bonjour!
>
> J'apprends le français récemment et J'ai installé myspell-fr pour
> "spelling check" dans gedit.  Quand je fais "spelling check" sur mon texte,
> plusieurs mots sont considérés invalides, tel que "j", "c", même "allez"!
> Je comprends que "j", "c" sont invalides parce que je utilise phrases
> <>,ect.  Mais pourquoi "allez" est invalide?
>
> Est-ce que vous avez une idée? Merci d'avance pour votre aide!
> Yuwen
>
>
> bonjour
> voici ce qui est installé sur mon ordi :
>
> dpkg -l |awk '/spell/ {print $1 "  "$2  "  "$3}'
>
>
> ii  aspell  0.60.7~20110707-4
> ii  aspell-en  2017.08.24-0-0.1
> ii  aspell-fr  0.50-3-8
> ii  dictionaries-common  1.27.2
> ii  enchant  1.6.0-11.1
> ii  hunspell  1.6.2-1
> ii  hunspell-en-us  20070829-7
> ii  hunspell-fr  1:6.1-1
> ii  hunspell-fr-classical  1:6.1-1
> ii  ifrench-gut  1:1.0-32
> ii  ispell  3.4.00-6
> ii  libaspell15:amd64  0.60.7~20110707-4
> ii  libenchant1c2a:amd64  1.6.0-11.1
> ii  libgtkspell0  2.0.16-1.1
> ii  libgtkspell3-3-0:amd64  3.0.9-1
> ii  libgtkspellmm-3.0-0v5:amd64  3.0.5+dfsg-1
> ii  libhfstospell9:amd64  0.4.5~r343-2.1
> ii  libhunspell-1.4-0:amd64  1.4.1-2+b2
> ii  libhunspell-1.6-0:amd64  1.6.2-1
> ii  libkf5sonnet5-data  5.37.0-2
> ii  libkf5sonnetcore5:amd64  5.37.0-2
> ii  libkf5sonnetui5:amd64  5.37.0-2
> ii  sonnet-plugins  5.37.0-2
>
> slt
> bernard
>

problème de dépendances avec bootcd

[bernard . schoenacker]

bonjour,

je n'arrive pas à installer bootcd sur une testing :



apt-cache depends bootcd
bootcd
  Dépend: init-system-helpers
  Dépend: genisoimage
  Dépend: cpio
cpio:i386
  Dépend: fdutils
  Dépend: file
file:i386
  Dépend: dosfstools
  Dépend: <<< paquet fautif
  Dépend: syslinux-common
  Dépend: initramfs-tools
 |Dépend: busybox
busybox-static
 |Dépend: busybox-static
 |Dépend: 
  Dépend: 
  Dépend: lsb-base
  Dépend: isolinux
 |Dépend: fdisk
fdisk:i386
  Dépend: util-linux
  Est en conflit avec: 
  Est en conflit avec: 
  Est en conflit avec: 
  Est en conflit avec: 
  Recommande: wodim
  Suggère: ssh
 |Suggère: discover
  Suggère: 
  Suggère: 
  Suggère: aufs-tools


ne serait ce point syslinux, mais comment corriger le tir ?

slt
bernard

"spelling check" pour le français

[Yuwen Dai]

Bonjour!

J'apprends le français récemment et J'ai installé myspell-fr pour "spelling
check" dans gedit.  Quand je fais "spelling check" sur mon texte, plusieurs
mots sont considérés invalides, tel que "j", "c", même "allez"!  Je
comprends que "j", "c" sont invalides parce que je utilise phrases
<>,ect.  Mais pourquoi "allez" est invalide?

Est-ce que vous avez une idée? Merci d'avance pour votre aide!
Yuwen

Re: rsyslog guardar log solo de clientes

[Juan Lavieri]

  
  



El 11-10-2017 a las 05:36 p.m.,
  Cristian Mitchell escribió:


  

  
El 11 oct. 2017 4:59 PM, "Alberto
  Cabrejas Pérez" 
  escribió:
  

  Tengo un server para gestionar log de estaciones
(rsyslog), funciona bien pero quisiera que los log
locales de ese servidor no los guardara, o sea que
solo almacene log remotos no los locales.
  
  -- 

  

  
Saludos,
  Alberto Cabrejas Pérez
  Administrador de Redes Informáticas
  ARTex S.A. Sucursal Granma
  http://www.scgr.artex.cu
  http://boletin.scgr.artex.cu
  Linux Usuario Registrado # 31 666
  Teléf.+53 23481956,23481912 (Ext 115)
  Atención al Cliente.+53 23483239
  Jabber: albe...@scgr.artex.cu
  "BE
FREE BE LINUX!!!"
  

  

  
  
  ARTex S.A. ...el arte de lo cubano. Este mensaje ha
  sido analizado por clamav antivirus y se considera
  totalmente
  limpio.
  

  


  

Redirigido a null
  


Simple y elegante Cristian.  ¡Excelente respuesta!

Saludos

-- 
Juan M Lavieri

Errar es de humanos, pero es mas humano culpar a los demás.
  

Re: x : keyboard not working

[Gene Heskett]

On Friday 06 October 2017 19:01:06 deloptes wrote:

>> Gene Heskett wrote:
> >> I have the same effect, but different log. I have way too many
> >> keyboards, none of which work.
>
>> Gene, do you have them connected at the same time? or you are
>> unplugging and plugging them after each other
>
>Only one at a time.

>> Just curious ...

>I found it via another post, console-setup was not installed, fixed that 
>and dpkg-reconfigured the keyboard and the console, rebooted, x is 
>running normally now. Thats a relief, I thought I was going to have to 
>hack on udev. :(

Updated 10/11/2017, one of my wheezy machines just did almost the exact 
same dance. Keyboard worked on the login screen though, but mouse was 
frozen.  Put new battery in mouse, powerdown rebooted, no effect. I 
could login but no mouse or keyboard to be found after x starts. Both of 
the above utils were installed, so I did the dpkg-reconfigure on both. 
That did not help, so that leaves the next most likely suspect.

The Dell Dimensions do NOT have a retaining clip on the far end of a pci 
card, and enough wiggle on the cable attached to the back plate end of a 
5i25, WILL cause it to work its way out of the socket by tipping up in 
front, which in turn causes all sorts of hate and discontent, with the 
present scenario one of the symptoms. IIRC I have some foamish stuff 
glued to the cover, obviously not enough pressure to hold it in solidly, 
yet...  So pull the cover tomorrow, and cobble up something that will 
hold it solidly into the socket.

OLd, off-lease Dell's are generally good stable machines, but that is the 
Dimensions Achilles heel.

>Thanks for asking.  To find the clue, install locate, updatedb, then 
>look 
>for README.Debian, and out of the 40 or 50 that spits out, read the one 
>with keyboard-config in its locate output line.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

Re: RE

[Jimmy Johnson]

On 10/11/2017 10:14 AM, Jape Person wrote:

On 10/11/2017 04:20 AM, debian-user@lists.debian.org wrote:

Check
Re: hplip and use of the "driver plugin"



I'm seeing something from debian-user@lists.debian.org that I've
never seen before. The most recent three days I have received a
single e-mail as quoted above.

Is this a normal function of the list? If so, what purpose does
it serve? The thread referenced died out in December of last year.

I just started getting back into the swing of things after a
prolonged medical issue, but didn't see any new messages in the
thread when I came back to check my mail. Is my attention being
drawn to the thread for some specific reason?

Thanks to anyone who can provide a clue!

JP


It looks like they are trying to get an update to: 
https://lists.debian.org/debian-user/2016/11/msg00769.html
Maybe a bug was filed on your behalf, anyways it looks like someone is 
interested in some kind of closure.  Just so you know you're not the 
only one, I had the same thing going on and gave them some feed back on 
a i965 video driver in stretch while in testing and I closed the bug 
cause it was working, I hope this helps. Also it shows that Debian is 
listening to you.

--
Jimmy Johnson

Ubuntu 14.04 LTS - KDE 4.13.2 - AMD A8-7600 - EXT4 at sda1
Registered Linux User #380263

Re: rsyslog guardar log solo de clientes

[Cristian Mitchell]

El 11 oct. 2017 4:59 PM, "Alberto Cabrejas Pérez" 
escribió:

Tengo un server para gestionar log de estaciones (rsyslog), funciona bien
pero quisiera que los log locales de ese servidor no los guardara, o sea
que solo almacene log remotos no los locales.
-- 

Saludos, *Alberto Cabrejas Pérez*
Administrador de Redes Informáticas
ARTex S.A. Sucursal Granma
http://www.scgr.artex.cu
http://boletin.scgr.artex.cu
Linux Usuario Registrado # 31 666
Teléf.+53 23481956,23481912 (Ext 115)
Atención al Cliente.+53 23483239
Jabber: albe...@scgr.artex.cu
"BE FREE BE LINUX!!!"

ARTex S.A. ...el arte de lo cubano. Este mensaje ha sido analizado por
clamav antivirus y se considera totalmente limpio.


Redirigido a null

Soccer Enthusiasts List

[Katrina Muller]

 Hi,

 Would you be interested in acquiring an email list of "Soccer Enthusiasts 
List" from USA?

 We also have data for Spa and Resort Visitors, Timeshare Owners, Entertainment 
Enthusiasts List, Skiers List, Golfers List, Basketball Enthusiasts, Baseball 
Enthusiasts, Students List, Apparel Buyers, Cycle Enthusiasts, Casino Players, 
Outdoor Enthusiasts, Boat Owners, Sucba Divers, Fishing Enthusiasts and many 
more..

 Each record in the list contains Contact Name (First, Middle and Last Name), 
Mailing Address, List type and Opt-in email address.

 All the contacts are opt-in verified, 100% permission based and can be used 
for unlimited multi-channel marketing.

 Please let me know your thoughts towards procuring the Soccer Enthusiasts List.

 Best Regards,
 Katrina Muller
 Research Analyst


 We respect your privacy, if you do not wish to receive any further emails from 
our end, please reply with a subject “Leave Out”.

rsyslog guardar log solo de clientes

[Alberto Cabrejas Pérez]

Tengo un server para gestionar log de estaciones (rsyslog), funciona 
bien pero quisiera que los log locales de ese servidor no los guardara, 
o sea que solo almacene log remotos no los locales.


--

Saludos, *Alberto Cabrejas Pérez*
Administrador de Redes Informáticas
ARTex S.A. Sucursal Granma
http://www.scgr.artex.cu
http://boletin.scgr.artex.cu
Linux Usuario Registrado # 31 666
Teléf.+53 23481956,23481912 (Ext 115)
Atención al Cliente.+53 23483239
Jabber: albe...@scgr.artex.cu
"BE FREE BE LINUX!!!"


ARTex S.A.   ...el arte de lo cubano. Este mensaje ha

sido analizado por clamav antivirus y se considera totalmente

limpio.

Re: What happened to libreoffice-common in archive?

[Sven Joachim]

On 2017-10-11 11:33 -0500, Nicholas Geovanis wrote:

> On Wed, Oct 11, 2017 at 10:24 AM, Sven Joachim  wrote:
>
>>
>> Since libreoffice takes 6-18 hours to build on the buildds, this happens
>> with every libreoffice upload. :-(
>>
>
> Since I don't understand all the implications of the bug report:
> Will it be the case that libreoffice (and apparently other packages)
> "eventually" makes it to the mirrors?

Yes (unless they FTBFS, that is).

> Or is special action on my part needed?

No, you don't have to do anything.

Cheers,
   Sven

Re: RE

[Jape Person]

On 10/11/2017 04:20 AM, debian-user@lists.debian.org wrote:
> Check 
> Re: hplip and use of the "driver plugin"
> 

I'm seeing something from debian-user@lists.debian.org that I've
never seen before. The most recent three days I have received a
single e-mail as quoted above.

Is this a normal function of the list? If so, what purpose does
it serve? The thread referenced died out in December of last year.

I just started getting back into the swing of things after a
prolonged medical issue, but didn't see any new messages in the
thread when I came back to check my mail. Is my attention being
drawn to the thread for some specific reason?

Thanks to anyone who can provide a clue!

JP

Re: AIDE defaults in debian stretch

[Dejan Jocic]

On 10-10-17, j...@bluemarble.net wrote:
> The Debian configuration files in AIDE on Debian seem to monitor a lot of
> files that I'm not sure need monitoring. Maybe someone could shed some
> light.
> 
> Is there a reason I should monitor /run? What about the /var/log/ files that
> are rotated. It often complains about that. How about systemd journal files?
> 
> Thanks.
> 
> 

I'm far from expert in this, just user of AIDE, so was hopping that
someone with more knowledge than me will shed some light on this.
Anyway, I did not like how AIDE works in Debian, looked overcomplicated
to me, so I've installed aide without recommends. If you do it like
that, you end up without aide-common package, which will make AIDE much
more vanilla like. You do not have any config file, nor cron job added
automatically. So, you need to do bit of learning that way and to
include in that aide.conf file what you want, and what you do not want.
Find some examples on net, like this one:

# define the path for creating the databases.
database=file:/var/lib/aide/aide.db
database_out=file:/var/lib/aide/aide.db.new
database_new=file:/var/lib/aide/aide.db.new
 
# define your own aide rule.
MYRULE =  p+n+u+g+s+m+c+xattrs+md5+sha1
 
# choose your directories/files you want in the database and which rule should 
be used.
/ MYRULE
 
# define your exceptions.
!/proc   # ignore /proc filesystem
!/sys# ignore /sys filesystem

That one is obvious overkill, because whole system will be checked
except /proc and /sys, but is good example how you can exclude what you
do not want to. Also, that one uses /var/lib/aide for databases, which
for sure is not recommended practice. Best practice would be to put
aide.conf, databases and even aide binary on, for example, USB that
would be inserted just for check. As for should you make AIDE check /run
and /var/log, not really sure. Some think that even some things under
/proc should be checked (not that AIDE can do it anyway). But checking
/var/log is annoying and bit of overkill, at least for me.

Hope that this helps you at least a bit.

Re: What happened to libreoffice-common in archive?

[Nicholas Geovanis]

On Wed, Oct 11, 2017 at 10:24 AM, Sven Joachim  wrote:

>
> Since libreoffice takes 6-18 hours to build on the buildds, this happens
> with every libreoffice upload. :-(
>

Since I don't understand all the implications of the bug report:
Will it be the case that libreoffice (and apparently other packages)
"eventually" makes it to the mirrors? Or is special action on
my part needed?
Thanks.Nick


> Cheers,
>Sven
>
>

Re: What happened to libreoffice-common in archive?

[Sven Joachim]

On 2017-10-11 16:11 +0200, Vincent Lefevre wrote:

> How can the following be possible?
>
> Note that I have not changed the mirror, and "apt update" does not
> give any error.
>
> cventin:~> apt-show-versions -a libreoffice-common
> libreoffice-common:all 1:5.4.2-1 install ok installed
> libreoffice-common:all 1:5.2.7-1 stable  ftp.fr.debian.org
> No stable-updates version
> libreoffice-common:all 1:5.4.1-1 testing ftp.fr.debian.org
> No unstable version
> No experimental version
> libreoffice-common:all 1:5.4.2-1 newer than version in archive
>  ^
>
> Why "No unstable version" while
>
>   https://packages.debian.org/sid/libreoffice-common
>
> says "libreoffice-common (1:5.4.2-1)"?
>
> Is this a bug with the mirror?

No, in dak: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=825398.

Basically dak is not ready for source-only uploads and considers any
arch: all package that has not become available at the same time as the
source package as obsolete.  If there are no reverse dependencies, the
arch: all package is removed.

Since libreoffice takes 6-18 hours to build on the buildds, this happens
with every libreoffice upload. :-(

Cheers,
   Sven

Re: What happened to libreoffice-common in archive?

[Cindy-Sue Causey]

On 10/11/17, Vincent Lefevre  wrote:
> How can the following be possible?
>
> Note that I have not changed the mirror, and "apt update" does not
> give any error.
>
> cventin:~> apt-show-versions -a libreoffice-common
> libreoffice-common:all 1:5.4.2-1 install ok installed
> libreoffice-common:all 1:5.2.7-1 stable  ftp.fr.debian.org
> No stable-updates version
> libreoffice-common:all 1:5.4.1-1 testing ftp.fr.debian.org
> No unstable version
> No experimental version
> libreoffice-common:all 1:5.4.2-1 newer than version in archive
>  ^
>
> Why "No unstable version" while
>
>   https://packages.debian.org/sid/libreoffice-common
>
> says "libreoffice-common (1:5.4.2-1)"?
>
> Is this a bug with the mirror?


Hi.. I had a (cognitively based) hard time understanding all this a
while back. Still kind of do (not fully grasp) primarily because I
haven't tracked down where to find the notes that hint at the "why".

The packages not showing are possibly sitting waiting, and in fact
this thread exists because you found that out. It has something to do
with Developers' design for *LOGICAL* reasons known to them.

It used to be that I would cherry pick my upgrades via "apt-get
install" after running "apt-get update"... because I'm on dialup
Internet access. I had to go that route to be able to upgrade AND stay
active doing poverty related advocacy shtuff, etc, online.

In going that route, I would install EVERYTHING while not knowing that
some things should NOT be upgraded even though a package does, yes,
sit ready in repositories.

I filed a couple bugs back when because trying to install some of
those packages led to "apt-get" UNINSTALLING other *critical*
packages, sometimes over 100 at a time. As it turned out, THAT kind of
thing is why those packages are on hold...

And I guess that's why they're showing up the way they do in your
"apt-show-versions" query feedback there. I'm using Buster/Testing,
and libreoffice-common is on hold... and has been on hold A VERY LONG
TIME. If I do "apt-get upgrade, I see something like this:

...
Reading package lists... Done
Building dependency tree
Reading state information... Done
Calculating upgrade... Done
The following packages have been kept back:
[list of ~140 well-known package names (including everything
libreoffice, cpp-6, cpp-7, yada]
...

NO complaints on packages of that size being held back, have I
mentioned dialup, yada-yada. *grin*

If you already do the following, then this is just noise, but I now go
the "apt-get upgrade" route after updating my apt.

In my special use case, I still "apt-get install" cherry pick after
the readout from the pending upgrade. The difference now is I cherry
pick from what UPGRADE would do versus where I used to pick from all
things shown by "apt-show-versions upgrade".

Oh, you know what, as I write that, I'm realizing that it sounds like
you're seeing what I could have used back then. It sounds like
apt-show-versions isn't reporting something that is in fact on hold.
*COOL!* If that used to exist even way back way, then I just wasn't
using the right flag to see that output or something.

Two things I scratch my head over are:

#1 Why do those packages show up at all, even though it's as being
"held back"? For now, I *a-sume* it's so that it's public knowledge
that there is the possibility for an upgrade if you just absolutely
must do so.

Knowing those packages are on hold is your sign to absolutely proceed
with caution, nay, just don't even go there. Although I
encountered something myself last night for the first time where it
wants me to upgrade some of developer packages. I may just do it since
it's because something else is not working right now anyway. Let's go
for total broke. *grin*

#2 WHY doesn't "apt-get install" balk when we try to manually install
those packages that on hold? I just quick tested, and it still
doesn't.

My rather unhumble is that apt-get should balk (not unlike
apt-show-versions there). It seems that apt-get should stop us and ask
us ARE WE SURE then specifically state that those packages have been
put on hold for intellectual reasons known to skilled Debian
Developers. After that, we can make an informed executive decision
regarding going ahead or canning the attempt for the time being..

Anyone have any feedback on what any other package managers do with
packages that are on hold? It would be about being a *VERY* newbie
friendly feature to at least warn before proceeding on...

Cindy :)
-- 
Cindy-Sue Causey
Talking Rock, Pickens County, Georgia, USA

* runs with... pumpkin spice cookies and a slushing cup of (cold) coffee *

What happened to libreoffice-common in archive?

[Vincent Lefevre]

How can the following be possible?

Note that I have not changed the mirror, and "apt update" does not
give any error.

cventin:~> apt-show-versions -a libreoffice-common
libreoffice-common:all 1:5.4.2-1 install ok installed
libreoffice-common:all 1:5.2.7-1 stable  ftp.fr.debian.org
No stable-updates version
libreoffice-common:all 1:5.4.1-1 testing ftp.fr.debian.org
No unstable version
No experimental version
libreoffice-common:all 1:5.4.2-1 newer than version in archive
 ^

Why "No unstable version" while

  https://packages.debian.org/sid/libreoffice-common

says "libreoffice-common (1:5.4.2-1)"?

Is this a bug with the mirror?

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

Re: Encriptant discs

[Iker]

Bones,

Gràcies per les aportaciones.

He vist que pots fer que la contrassenya LUKS la llegeixi d'un medi 
extern. D'aquesta manera evito haver d'introduïr la contrassenya 
(keylogger) i tinc un dispositiu físic (clau) per controlar l'encesa.


Moltes gràcies,

*Iker.*



On 11/10/17 14:18, Narcis Garcia wrote:

La instal·lació senzilla implica establir una partició «oberta» per a
/boot (on hi haurà el nucli) i instal·lar el gestor d'arrencada (GRUB) a
la unitat interna de l'ordinador (disc dur).
Això vol dir, però, que seria possible que un expert amb accés físic a
l'ordinador aturat, manipulés aquesta arrencada per a enregistrar la
contrasenya que teclegis la propera vegada, i per això hauries d'iniciar
tu l'ordinador després sense saber d'aquesta manipulació.
No és una situació gaire probable.

La manera més fàcil que jo he vist per evitar aquesta possibilitat és la
memòria USB per arrencar. Per instal·lar disposes de:
- El mitjà d'inici i instal·lació (Debian CD/DVD/USB)
- La unitat interna de l'ordinaodr (disc dur on anirà el sistema)
- Una memòria USB que ja connectes per començar.

A l'etapa del particionat, li dius que ho vols fer manualment, a la
unitat interna hi estableixes una sola partició per xifrat i a la
memòria USB una partició per a /boot
Al final del procés d'instal·lació, li dius que instal·li el gestor
d'arrencada a la memòria USB.

Res més; es tractarà d'iniciar amb la memòria USB i sempre que marxis te
l'emportes amb tu. Això et farà confiar amb què l'arrencada «oberta» no
haurà estat mai manipulada.


El 11/10/17 a les 14:00, Pedro ha escrit:

L'atac que pretén efectuar-se a través de grub i kernel se li diu evil
maid attack [1] [2]

Que jo sàpiga no pots xifrar grub i kernel. Però pots instal·lar-lo en
un usb que portaràs amb el teu clauer, i que haurà d'estar posat només
en el moment d'inici del sistema: conec una persona que ho fa.

També pots aprofitar per posar en aquest usb el header de luks
d'aquesta manera el teu ordinador "no tindrà res identificable" per
demostrar que està xifrat.

això és lo que es pot fer, però mai ho he fet, no sé com es fa

si t'hi fiques amb això i fas unes notes, comparteix; si algú té
informació de com es fa li agrairia molt

[1] info: https://en.wikipedia.org/wiki/Rootkit#Bootkits
[2] Automated Linux evil maid attack: https://github.com/GDSSecurity/EvilAbigail

2017-10-11 12:54 GMT+02:00 Iker :

Bones,

 Estic endinsant-me en el xifrat de discos i m'adono que en molts casos
només es xifra(LUKS) el sistema de /root mentre que altres parlen de xifrat
arranc(grub) i kernel (boot).

 Suposant que la contrassenya de LUKS l'introdueix l'usuari a l'arranc i
que el kernel és un kernel genèric (no té res afegit que m'interessi amagar)
no sé veure de quina forma es pot esquivar el xifrat del sistema a travès de
manipular el grub o el kernel. Entenc que podries evitar que s'arranquès
però mai accedir a les dades xifrades. Quan, a travès del gestor
d'instal·lació de Debian, trio xifrar el disc no estic xifrant grub ni
kernel, oi?.

 La pregunta és: xifrar grub i kernel dóna alguna seguretat extra perque
poden ser vector d'intrusió?

 Gràcies,

Iker.

Re: Encriptant discs

[Narcis Garcia]

La instal·lació senzilla implica establir una partició «oberta» per a
/boot (on hi haurà el nucli) i instal·lar el gestor d'arrencada (GRUB) a
la unitat interna de l'ordinador (disc dur).
Això vol dir, però, que seria possible que un expert amb accés físic a
l'ordinador aturat, manipulés aquesta arrencada per a enregistrar la
contrasenya que teclegis la propera vegada, i per això hauries d'iniciar
tu l'ordinador després sense saber d'aquesta manipulació.
No és una situació gaire probable.

La manera més fàcil que jo he vist per evitar aquesta possibilitat és la
memòria USB per arrencar. Per instal·lar disposes de:
- El mitjà d'inici i instal·lació (Debian CD/DVD/USB)
- La unitat interna de l'ordinaodr (disc dur on anirà el sistema)
- Una memòria USB que ja connectes per començar.

A l'etapa del particionat, li dius que ho vols fer manualment, a la
unitat interna hi estableixes una sola partició per xifrat i a la
memòria USB una partició per a /boot
Al final del procés d'instal·lació, li dius que instal·li el gestor
d'arrencada a la memòria USB.

Res més; es tractarà d'iniciar amb la memòria USB i sempre que marxis te
l'emportes amb tu. Això et farà confiar amb què l'arrencada «oberta» no
haurà estat mai manipulada.


El 11/10/17 a les 14:00, Pedro ha escrit:
> L'atac que pretén efectuar-se a través de grub i kernel se li diu evil
> maid attack [1] [2]
> 
> Que jo sàpiga no pots xifrar grub i kernel. Però pots instal·lar-lo en
> un usb que portaràs amb el teu clauer, i que haurà d'estar posat només
> en el moment d'inici del sistema: conec una persona que ho fa.
> 
> També pots aprofitar per posar en aquest usb el header de luks
> d'aquesta manera el teu ordinador "no tindrà res identificable" per
> demostrar que està xifrat.
> 
> això és lo que es pot fer, però mai ho he fet, no sé com es fa
> 
> si t'hi fiques amb això i fas unes notes, comparteix; si algú té
> informació de com es fa li agrairia molt
> 
> [1] info: https://en.wikipedia.org/wiki/Rootkit#Bootkits
> [2] Automated Linux evil maid attack: 
> https://github.com/GDSSecurity/EvilAbigail
> 
> 2017-10-11 12:54 GMT+02:00 Iker :
>> Bones,
>>
>> Estic endinsant-me en el xifrat de discos i m'adono que en molts casos
>> només es xifra(LUKS) el sistema de /root mentre que altres parlen de xifrat
>> arranc(grub) i kernel (boot).
>>
>> Suposant que la contrassenya de LUKS l'introdueix l'usuari a l'arranc i
>> que el kernel és un kernel genèric (no té res afegit que m'interessi amagar)
>> no sé veure de quina forma es pot esquivar el xifrat del sistema a travès de
>> manipular el grub o el kernel. Entenc que podries evitar que s'arranquès
>> però mai accedir a les dades xifrades. Quan, a travès del gestor
>> d'instal·lació de Debian, trio xifrar el disc no estic xifrant grub ni
>> kernel, oi?.
>>
>> La pregunta és: xifrar grub i kernel dóna alguna seguretat extra perque
>> poden ser vector d'intrusió?
>>
>> Gràcies,
>>
>> Iker.
> 

Re: Encriptant discs

[Pedro]

L'atac que pretén efectuar-se a través de grub i kernel se li diu evil
maid attack [1] [2]

Que jo sàpiga no pots xifrar grub i kernel. Però pots instal·lar-lo en
un usb que portaràs amb el teu clauer, i que haurà d'estar posat només
en el moment d'inici del sistema: conec una persona que ho fa.

També pots aprofitar per posar en aquest usb el header de luks
d'aquesta manera el teu ordinador "no tindrà res identificable" per
demostrar que està xifrat.

això és lo que es pot fer, però mai ho he fet, no sé com es fa

si t'hi fiques amb això i fas unes notes, comparteix; si algú té
informació de com es fa li agrairia molt

[1] info: https://en.wikipedia.org/wiki/Rootkit#Bootkits
[2] Automated Linux evil maid attack: https://github.com/GDSSecurity/EvilAbigail

2017-10-11 12:54 GMT+02:00 Iker :
> Bones,
>
> Estic endinsant-me en el xifrat de discos i m'adono que en molts casos
> només es xifra(LUKS) el sistema de /root mentre que altres parlen de xifrat
> arranc(grub) i kernel (boot).
>
> Suposant que la contrassenya de LUKS l'introdueix l'usuari a l'arranc i
> que el kernel és un kernel genèric (no té res afegit que m'interessi amagar)
> no sé veure de quina forma es pot esquivar el xifrat del sistema a travès de
> manipular el grub o el kernel. Entenc que podries evitar que s'arranquès
> però mai accedir a les dades xifrades. Quan, a travès del gestor
> d'instal·lació de Debian, trio xifrar el disc no estic xifrant grub ni
> kernel, oi?.
>
> La pregunta és: xifrar grub i kernel dóna alguna seguretat extra perque
> poden ser vector d'intrusió?
>
> Gràcies,
>
> Iker.

Encriptant discs

[Iker]

Bones,

Estic endinsant-me en el xifrat de discos i m'adono que en molts 
casos només es xifra(LUKS) el sistema de */root *mentre que altres 
parlen de xifrat *arranc(grub*) i *kernel (boot)*.


Suposant que la contrassenya de LUKS l'introdueix l'usuari a 
l'arranc i que el kernel és un kernel genèric (no té res afegit que 
m'interessi amagar) no sé veure de quina forma es pot esquivar el xifrat 
del sistema a travès de manipular el *grub* o el *kernel*. Entenc que 
podries evitar que s'arranquès però mai accedir a les dades xifrades. 
Quan, a travès del gestor d'instal·lació de Debian, trio xifrar el disc 
no estic xifrant grub ni kernel, oi?.


La pregunta és: xifrar grub i kernel dóna alguna seguretat extra 
perque poden ser vector d'intrusió?


Gràcies,

*Iker.*

Re: Error de instalacion e inicio

[Guillermo Cordeiro]

On 17-10-10 13:50:59, Alejandro Franco Trujillo Ayala wrote:
>Buenas tades

No se como colocar este fallo que aparece en mi computadora cada vez
que inicia:

   ACPI Error: No handler for Region [EC__] (9583ee0e7870)
   [EmbeddedControl] (20160831/evregion-166)

   ACPI Error: Region EmbeddedControl (ID=3) has no handler
   (20160831/exfldio-299)

   ACPI Error: Method parse/execution failed [\_SB.PCI0.LPCB.EC._REG]
   (Node 9583ee0e8bb8), AE_NOT_EXIST (20160831/psparse-543)

   nouveau :01:00.0: unknown chipset (136000a1)

   i915 :00:02.0: firmware: failed to load i915/skl_dmc_ver1_26.bin
   (-2)

   ath10k_pci :3e:00.0: firmware: failed to load ath10k/pre-cal-
   pci-:3e:00.0.bin (-2)

   ath10k_pci :3e:00.0: firmware: failed to load ath10k/cal-pci-
   :3e:00.0.bin (-2)

   ath10k_pci :3e:00.0: firmware: failed to load
   ath10k/QCA6174/hw3.0/firmware-5.bin (-2)

   ath10k_pci :3e:00.0: could not fetch firmware file
   'ath10k/QCA6174/hw3.0/firmware-5.bin': -2

intente instalar los firmware y no arregla nada

a los mismo con bumblebee, mi laptop ya no enciende al instalarlo

y al parecer el hdmi no funciona para doble pantalla


Mi laptop es una MSI gs73vr 6rf
procesador i7 HQ

GPU intel 530 /nvidia GTX 1060 (hibrido)

ssd 250GB y hdd 1TB


gracias espero ayude y si se puede una solucion al problemas

siguan adelante!


Me sucede a veces y un "workaround" es deshabilitar el KMS (Kernel Mode
Setting) para ese módulo. Esto lo puedes hacer de dos formas:

   * En el GrUB, añadiendo en la línea del kernel "nouveau.modeset=0",
   ya sea durante el arranque o en /etc/default/grub
   * Añadiende un fichero de configuración para modprobe, por ejemplo
   /etc/modprobe.d/disable-modeset-nouveau.conf con el contenido:
   options nouveau modeset=0

Un saludo


signature.asc
Description: PGP signature

Error de instalacion e inicio

[Alejandro Franco Trujillo Ayala]

Buenas tades
No se como colocar este fallo que aparece en mi computadora cada vez
que inicia:

ACPI Error: No handler for Region [EC__] (9583ee0e7870)
[EmbeddedControl] (20160831/evregion-166)

ACPI Error: Region EmbeddedControl (ID=3) has no handler
(20160831/exfldio-299)

ACPI Error: Method parse/execution failed [\_SB.PCI0.LPCB.EC._REG]
(Node 9583ee0e8bb8), AE_NOT_EXIST (20160831/psparse-543)

nouveau :01:00.0: unknown chipset (136000a1)

i915 :00:02.0: firmware: failed to load i915/skl_dmc_ver1_26.bin
(-2)

ath10k_pci :3e:00.0: firmware: failed to load ath10k/pre-cal-
pci-:3e:00.0.bin (-2)

ath10k_pci :3e:00.0: firmware: failed to load ath10k/cal-pci-
:3e:00.0.bin (-2)

ath10k_pci :3e:00.0: firmware: failed to load
ath10k/QCA6174/hw3.0/firmware-5.bin (-2)

ath10k_pci :3e:00.0: could not fetch firmware file
'ath10k/QCA6174/hw3.0/firmware-5.bin': -2

intente instalar los firmware y no arregla nada 

a los mismo con bumblebee, mi laptop ya no enciende al instalarlo

y al parecer el hdmi no funciona para doble pantalla 


Mi laptop es una MSI gs73vr 6rf 
procesador i7 HQ 

GPU intel 530 /nvidia GTX 1060 (hibrido)

ssd 250GB y hdd 1TB


gracias espero ayude y si se puede una solucion al problemas

siguan adelante!

atte.

Alejandro Trujillo 

Re: sed et travail sur un bloc contenant un motif

[Migrec]

Le 10/10/2017 à 08:57, Dominique Dumont a écrit :

On Monday, 9 October 2017 22:11:31 CEST Alexandre Hoïde wrote:

  sed, peut-être… perso je ne vois pas, mais peut-être. Par contre, gawk ?

Pas d'idée pour sed non plus.

L'exemple avec awk m'a fait penser à une option peu connue de perl qui permet
de séparer un fichier en paragraphes ( -000) . Ce qui donne avec l'exemple de
fichier ldif donné par Alexandre:

$ perl -n -000 -e 'print if /sAMAccountName: smith/;' fichier.ldif
dn: CN=John Smith,OU=Users,DC=contoso,DC=com
proxyAddresses: SMTP:sm...@contoso.com
proxyAddresses: smtp:john.sm...@contoso.com
proxyAddresses: smtp:jsm...@elsewhere.com
proxyAddresses: MS:ORG/ORGEXCH/JOHNSMITH
sAMAccountName: smith

dn: CN=John Smith,OU=Users,DC=contoso,DC=com
proxyAddresses: SMTP:sm...@contoso.com
proxyAddresses: smtp:john.sm...@contoso.com
proxyAddresses: smtp:jsm...@elsewhere.com
proxyAddresses: MS:ORG/ORGEXCH/JOHNSMITH
sAMAccountName: smith


Merci, je vais tester ça pour voir si ça va bien mais ça semble pas mal. J'ai 
un fichier texte de 150Mo à parser donc je choisirai l'option la plus rapide !

J'ai eu ceci également ($1 motif, $2 fichier ldif à parser:

#!/bin/bash

# Recherche d'un uid
sed -n "/^dn: uid=/{:a;N;/\n$/!ba;/$1/p}" $2

Memtest86+ pas utilisable

[kaliderus]

Le bonjour la liste,

Alors voilà je souhaite tester la RAM de mon portable (un Thinkpad
X230), et j'ai donc installé Memtest86+ (version livrée avec Jessie).

La ligne (en fait plusieurs avec différentes options) apparaît bien
dans le menu Grub, et quand je la sélectionne (j'ai testé chaque
option), j'ai droit à :
Un écran noir, le ventilateur de la machine accélère et ...rien de rien ...

Même comportement avec Memtest86.

Forcément le résultat n'est pas très probant.

C'est comme si l'utilitaire se lançait mais sans passer par la carte
graphique (qui fonctionne très bien, je l'utilise 10h/jour).

J'ai beau jouer avec les options de grub en ligne de commande, changer
le background_color, vérifier la présence des fichiers de Memtest86+,
demander à glouglou ce qu'il en pense etc ... toujours rien de rien,
tout semble installé comme il se doit.

Je me demande même par quel bout prendre le problème, hormis remonter
l'anomalie aux développeurs.

Si une idée lumineuse se présente je suis preneur.

Merci.

ps : La dernière version bootable sur clé USB fonctionne, mais ce
n'est pas mon besoin.

Re: créer live cd d'un système existant

[jérémy prego]

Le 10/10/2017 à 10:28, Raphaël POITEVIN a écrit :

jérémy prego  writes:

je dispose d'un système existant que j'aimerai convertir en live cd,
après y avoir apporté toutes les modifications nécessaires.

j'ai peut être mal cherché, mais j'ai pas trouvé de tuto en français
pour réaliser cette opération... Par contre, pour modifier un livecd
existant ça oui, mais dans l'idéale j'aimerai convertir mon système en
live cd ou live dvd bien sûre, selon la taille que ça fera ...

Regarde de ce côté :
http://www.linuxpedia.fr/doku.php/debian/debian_fabriquer_un_cd_live

super, ça fonctionne vraiment bien !

Raphaël


jerem