Re: bash completion and spaces

[Victor Sudakov]

David Wright wrote:
> > > > 
> > > > I have an example app which can be run only as "app3 -h test1 -s foo" or
> > > > "app3 -h test2 -s bar". So I decided to provide it with a small manual
> > > > completion for convenience.
> > > > 
> > > > [vas@test2 ~]$ ./list.sh 
> > > > -h test1 -s foo
> > > > -h test2 -s bar
> > > > [vas@test2 ~]$ complete -C ./list.sh app3
> > > > [vas@test2 ~]$
> > > > 
> > > > The result however is discouraging, the completion mechanism won't add
> > > > whole lines of parameters, it's trying to split on spaces (here I press
> > > >  several times:
> > > > 
> > > > [vas@test2 ~]$ app3 -h -h -h test
> > > > 
> > > > Can you please give a hint how to make it complete "app3" with either
> > > > "-h test1 -s foo" or "-h test2 -s bar" as a whole?
> > > > 
> > > > I would not like to make all this too complicated, write complex
> > > > completion funcions if possible. A static (-W) completion would be even
> > > > better.
> > > 
> > > Perhaps:
> > > 
> > > alias app3a='app3 -h test1 -s foo'
> > > alias app3b='app3 -h test2 -s bar'
> > 
> > No, not alias, I'd like to do it via bash completion. I may want to make
> > it dynamic eventually, the problem is in the spaces.
> 
> It's claimed that you can overcome this, in
> 
> https://stackoverflow.com/questions/11070370/bash-completion-for-strings-with-spaces-and-punctuation

Yes, this describes my case, unfortunately it lacks an example. Maybe
googling for COMP_WORDBREAKS will yield some examples. Thanks for the
link.

> 
> but I'd be balancing the possibility of side-effects with redefining Space
> against the benefits of completing a few strings.

If COMP_WORDBREAKS cannot be redefined within one and only one
completion, then perhaps the game is not worth the candle.

BTW on my current Debian system I don't see the space character in 
$COMP_WORDBREAKS.

> 
> I also don't understand what you mean by "dynamic", particularly as
> you wrote "A static (-W) completion" above. 

I actually meant that a simple example of static completion with spaces
(if someone cares to provide it) would help me understand the idea, and
I could use it in a dynamic completion later if I need to.

> You could always generate
> lists of aliases dynamically, too.

Well, I can make bash regenerate a completion list every time I
enter "app3", probably it can be done with aliases as well?

-- 
Victor Sudakov VAS4-RIPE
http://vas.tomsk.ru/
2:5005/49@fidonet


signature.asc
Description: PGP signature

Re: No networking after resume from suspend

[Greg Wooledge]

On Fri, Apr 23, 2021 at 09:59:09PM -0500, David Wright wrote:
> OTOH my startup files set Mywiredifname for scripts to use, where:
> 
> Mywiredifname=$(ip -o link show | sed -e '/^[0-9]\+: [^e]/d;s/[0-9]\+: 
> \([^:]\+\): .*/\1/;q')
> 
> $ echo $Mywiredifname 
> enp3s0

unicorn:~$ ip -o link show | awk -F': ' '{print $2}'
lo
lan0

You're relying on the "default" names which ensure that an ethernet
interface begins with an 'e', so we could fold that in as well:

ip -o link show | awk -F': ' '$2 ~ /^e/ {print $2}'

It's slightly prettier than yours, and doesn't rely on GNUisms (which are
probably not a big issue since a system with "ip" probably has GNU sed).

> (My days of running multiple ethernet cards are long gone,
> so sed will quit after one match.)

Sounds like a naive assumption.  Some motherboards have dual NICs built in,
don't they?

Re: bash completion and spaces

[David Wright]

On Thu 22 Apr 2021 at 21:38:53 (+0700), Victor Sudakov wrote:
> David Wright wrote:
> > Victor Sudakov wrote:
> > > 
> > > I have an example app which can be run only as "app3 -h test1 -s foo" or
> > > "app3 -h test2 -s bar". So I decided to provide it with a small manual
> > > completion for convenience.
> > > 
> > > [vas@test2 ~]$ ./list.sh 
> > > -h test1 -s foo
> > > -h test2 -s bar
> > > [vas@test2 ~]$ complete -C ./list.sh app3
> > > [vas@test2 ~]$
> > > 
> > > The result however is discouraging, the completion mechanism won't add
> > > whole lines of parameters, it's trying to split on spaces (here I press
> > >  several times:
> > > 
> > > [vas@test2 ~]$ app3 -h -h -h test
> > > 
> > > Can you please give a hint how to make it complete "app3" with either
> > > "-h test1 -s foo" or "-h test2 -s bar" as a whole?
> > > 
> > > I would not like to make all this too complicated, write complex
> > > completion funcions if possible. A static (-W) completion would be even
> > > better.
> > 
> > Perhaps:
> > 
> > alias app3a='app3 -h test1 -s foo'
> > alias app3b='app3 -h test2 -s bar'
> 
> No, not alias, I'd like to do it via bash completion. I may want to make
> it dynamic eventually, the problem is in the spaces.

It's claimed that you can overcome this, in

https://stackoverflow.com/questions/11070370/bash-completion-for-strings-with-spaces-and-punctuation

but I'd be balancing the possibility of side-effects with redefining Space
against the benefits of completing a few strings.

I also don't understand what you mean by "dynamic", particularly as
you wrote "A static (-W) completion" above. You could always generate
lists of aliases dynamically, too.

Cheers,
David.

Re: No networking after resume from suspend

[David Wright]

On Fri 23 Apr 2021 at 13:23:31 (-0400), Cindy Sue Causey wrote:
> 
> All of the adapters that come as a single item have worked great. You
> just have to track down the new identifier, e.g. one of mine is
> "enx00909e9dd1ee". That long value goes wherever one normally types in
> eth0, enp1s0, eno1, etc.
> 
> Dmesg often reports that value. I grep /var/log/kern* for the word
> "renamed" because that's what I always remember first. Except that
> then there's having to open kern.log in a text editor occasionally so
> yeah, dmesg rocks if one can remember to go that route.
> 
> Grep works with dmesg, too. I just tested it. Only one line to read as
> output. Very cognitively friendly! :)

That seems like a lot of typing. You should be able to read your
interface name with 5 keystrokes:

$ ip a

and l(ink) will work in place of a(ddress).

OTOH my startup files set Mywiredifname for scripts to use, where:

Mywiredifname=$(ip -o link show | sed -e '/^[0-9]\+: [^e]/d;s/[0-9]\+: 
\([^:]\+\): .*/\1/;q')

$ echo $Mywiredifname 
enp3s0
$ 

(My days of running multiple ethernet cards are long gone,
so sed will quit after one match.)

Cheers,
David.

Re: OT: Freestanding spreadsheet program?

[Weaver]

On 24-04-2021 08:25, Stefan Monnier wrote:
>> Back in the good (bad?) old days of TRS-80, all we had was VisiCalc. Simple.
>> Today, is there a useful spreadsheet program that does not rely on all the
>> baggage associated with either an "office suite," or
>> a "desktop environment?"
> 
> I can mention `gnumeric` and if you're into Emacs I can also also
> suggest SES.

Plus 1 for gnumeric.
Excellent.
Doesn't drop data on large data sets as much vaunted proprietary
programmes do.
Cheers!

Harry.

Re: OT: Freestanding spreadsheet program?

[Stefan Monnier]

> Back in the good (bad?) old days of TRS-80, all we had was VisiCalc. Simple.
> Today, is there a useful spreadsheet program that does not rely on all the
> baggage associated with either an "office suite," or
> a "desktop environment?"

I can mention `gnumeric` and if you're into Emacs I can also also
suggest SES.


Stefan

Re: OT: Freestanding spreadsheet program?

[Jude DaShiell]

sure, teapot.


On Fri, 23 Apr 2021, Dan Ritter wrote:

> Bob Bernstein wrote:
> > Back in the good (bad?) old days of TRS-80, all we had was VisiCalc. Simple.
> >
> > Today, is there a useful spreadsheet program that does not rely on all the
> > baggage associated with either an "office suite," or a "desktop
> > environment?"
> >
>
> Sure.
>
> treesheets
>
> pyspread
>
> sc
>
> -dsr-
>
>

Re: OT: Freestanding spreadsheet program?

[Dan Ritter]

Bob Bernstein wrote: 
> Back in the good (bad?) old days of TRS-80, all we had was VisiCalc. Simple.
> 
> Today, is there a useful spreadsheet program that does not rely on all the
> baggage associated with either an "office suite," or a "desktop
> environment?"
> 

Sure.

treesheets

pyspread

sc

-dsr-

OT: Freestanding spreadsheet program?

[Bob Bernstein]

Back in the good (bad?) old days of TRS-80, all we had was 
VisiCalc. Simple.


Today, is there a useful spreadsheet program that does not rely 
on all the baggage associated with either an "office suite," or a 
"desktop environment?"


Thx,


--
"...that there is no getting away from the central position of the individual I believe 
anyone can see for himself merely by observing that the individual does not regard the 
following to be a senseless question: "Under what conditions would you draw the 
conclusion that everyone in the world except yourself had gone crazy?"

Bridgman _The Nature of Physical Theory_ (1936)

Re: installing two versions of a user application

[Morgan Read]

On 23/04/2021 6:20 pm, Stefan Monnier wrote:
> I suggest you look inside those .deb files: they seem to both name the
> package they install `master-pdf-editor` but dpkg does not allow
> installing different versions of a given package at the same time.
> So you'll need to edit those `.deb` to change the package name from
> `master-pdf-editor` to (say) `master-pdf-editor4` and
> `master-pdf-editor5`, for example.

Thanks Stefan, I was wondering about something along those lines, but...
Inside master-pdf-editor-4.3.89_qt5.x86_64.deb I find:
data.tar.xz > /./opt/master-pdf-editor-4/masterpdfeditor4
  /./usr/share/icons/hicolor/128x128/apps/masterpdfeditor4.png
  /./usr/share/applications/masterpdfeditor4.desktop

And, inside master-pdf-editor-5.7.08-qt5_included.x86_64.deb I find:
data.tar.xz > /./opt/master-pdf-editor-5/masterpdfeditor5.sh
  /./opt/master-pdf-editor-5/masterpdfeditor5.png
  /./opt/master-pdf-editor-5/masterpdfeditor5
  /./usr/share/applications/masterpdfeditor5.desktop
  /./usr/share/icons/hicolor/256x256/apps/masterpdfeditor5.png

So, everything looks fairly well distinguished - but, I know *nothing*
about packaging...

Thanks

<>

signature.asc
Description: OpenPGP digital signature

Xen VGA passthrough working again in Debian stable!

[Fr. Chuck Zmudzinski, C.P.M.]

An old bug (#776742) reported over six years ago by Brian Paterni 
concerns Xen and VGA passthrough for full graphics performance in a 
virtual machine, and it was recently closed without being solved. About 
seven years ago, Xen had the best support for this feature but it has 
broken for a long time in Xen and that is why most users who want the 
VGA passthrough feature on their workstations are now using KVM/virtio 
based solutions instead of Xen/xl. However, I never gave up on Xen and 
with some workarounds I have been able to do VGA passthrough of an Intel 
integrated graphics device (IGD) using Xen and Debian but not since the 
days of wheezy have I been able to do this with unpatched and official 
up-to-date Debian software… until now!


The magic fix came with the backport of Qemu 5.2 to buster. After 
installing it on my system, I was able to fairly easily configure my 
Debian system to passthrough the Intel IGD. Thank you Debian Qemu Team 
and Debian Xen Team!


How I configured it:

Hardware: Haswell core i5-4590S with Intel HD Graphics 4600 and ASRock 
B85M-Pro4 Motherboard with 16 GB memory


Software: Up-to-date amd64 Debian buster using the Xen metapackage which 
installs some Qemu 3.1 packages. Then I installed the Qemu 5.2 backport 
of Qemu and its dependencies from buster-backports.


Passthrough configuration: Default bios (seabios), default device model, 
default toolstack (xl), but a non-default setting for rdm in xl.cfg:


rdm = "strategy=host,policy=relaxed"

This rdm setting might not be needed for newer Intel IGD devices. Mine 
is about seven years old.

n
To make the IGD assignable, in Xen it is very simple, no need to 
blacklist drivers (assuming the Intel IGD is not running an X or Wayland 
display):


sudo xl pci-assignable-add :02.0

In /etc/default/grub I have:

GRUB_CMDLINE_XEN="dom0_mem=2G,max:2G smt=false pv-l1tf=false iommu=1"

Not all these settings in grub are necessary, but this is a setting to 
tweak, as well as the GRUB_CMDLINE_LINUX_DEFAULT setting in 
/etc/default/grub if it doesn’t work on the first try.


Windows 10 works great with this setup – I also passed through the USB 
3.0 controller and the onboard realtekHD audio device and do not 
experience any crackling sounds or any problems that are sometimes 
reported when using KVM. Also, Xen’s GPLPV drivers installed fairly 
easily for accelerated disk and network performance.


A Debian buster virtual machine did not work, because the 4.19 kernel 
did not successfully configure the passed through hardware with default 
settings. I will try a newer backported kernel and see if that works and 
file a bug report about that if it isn't fixed in the latest kernels.


Going forward, I think Xen might be worth a try again for the difficult 
problem of getting VGA passthrough to Windows working on Linux. I think 
even the infamous code 43 error that Nvidia GeForce cards trigger is 
being fixed by Nvidia so the fact that Xen does not have a workaround 
for that problem like KVM does is not a deal breaker for Xen even for 
Nvidia users.

Xen VGA passthrough working again in Debian stable!

[Chuck Zmudzinski]

An old bug (#776742) reported over six years ago by Brian Paterni 
concerns Xen and VGA passthrough for full graphics performance in a 
virtual machine, and it was recently closed without being solved. About 
seven years ago, Xen had the best support for this feature but it has 
broken for a long time in Xen and that is why most users who want the 
VGA passthrough feature on their workstations are now using KVM/virtio 
based solutions instead of Xen/xl. However, I never gave up on Xen and 
with some workarounds I have been able to do VGA passthrough of an Intel 
integrated graphics device (IGD) using Xen and Debian but not since the 
days of wheezy have I been able to do this with unpatched and official 
up-to-date Debian software… until now!


The magic fix came with the backport of Qemu 5.2 to buster. After 
installing it on my system, I was able to fairly easily configure my 
Debian system to passthrough the Intel IGD. Thank you Debian Qemu Team 
and Debian Xen Team!


How I configured it:

Hardware: Haswell core i5-4590S with Intel HD Graphics 4600 and ASRock 
B85M-Pro4 Motherboard with 16 GB memory


Software: Up-to-date amd64 Debian buster using the Xen metapackage which 
installs some Qemu 3.1 packages. Then I installed the Qemu 5.2 backport 
of Qemu and its dependencies from buster-backports.


Passthrough configuration: Default bios (seabios), default device model, 
default toolstack (xl), but a non-default setting for rdm in xl.cfg:


rdm = "strategy=host,policy=relaxed"

This rdm setting might not be needed for newer Intel IGD devices. Mine 
is about seven years old.

n
To make the IGD assignable, in Xen it is very simple, no need to 
blacklist drivers (assuming the Intel IGD is not running an X or Wayland 
display):


sudo xl pci-assignable-add :02.0

In /etc/default/grub I have:

GRUB_CMDLINE_XEN="dom0_mem=2G,max:2G smt=false pv-l1tf=false iommu=1"

Not all these settings in grub are necessary, but this is a setting to 
tweak, as well as the GRUB_CMDLINE_LINUX_DEFAULT setting in 
/etc/default/grub if it doesn’t work on the first try.


Windows 10 works great with this setup – I also passed through the USB 
3.0 controller and the onboard realtekHD audio device and do not 
experience any crackling sounds or any problems that are sometimes 
reported when using KVM. Also, Xen’s GPLPV drivers installed fairly 
easily for accelerated disk and network performance.


A Debian buster virtual machine did not work, because the 4.19 kernel 
did not successfully configure the passed through hardware with default 
settings. I will try a newer backported kernel and see if that works and 
file a bug report about that if it isn't fixed in the latest kernels.


Going forward, I think Xen might be worth a try again for the difficult 
problem of getting VGA passthrough to Windows working on Linux. I think 
even the infamous code 43 error that Nvidia GeForce cards trigger is 
being fixed by Nvidia so the fact that Xen does not have a workaround 
for that problem like KVM does is not a deal breaker for Xen even for 
Nvidia users.

Re: Nvidia graphic card and newer kernels (open source driver)

[Felix Miata]

Malte Marwedel composed on 2021-04-23 18:31 (UTC+0200):

> Any Ideas? I don't want to install the closed source module, I had 
> enough pain with fglrx ~10 years ago.

I've never installed NVidia's proprietary drivers on any PC I own. I did do an
install once too many moons ago to remember on someone else's NVIDIA G98 
[GeForce
8400 GS Rev. 2] (which since became mine).

These are the NVidia cards I have in current use, all running on the FOSS 
default
DIX driver, Modesetting:

NVIDIA C61 [GeForce 6150SE nForce 430]
NVIDIA G84 [GeForce 8600 GT]
NVIDIA G98 [GeForce 8400 GS Rev. 2]
NVIDIA GF108 [GeForce GT 630]
NVIDIA GF119 [NVS 310]
NVIDIA GT218 [GeForce 210]

The FOSS X drivers from upstream are in source packages named xf86-video-.
Debian releases them compiled in the form xserver-xorg-video-. They are
called DDX as shorthand for device dependent X. The Modesetting DIX is newer
technology, provided by the X server package itself, so is not optional. DIX is
shorthand for device independent X. It supports AMD, ATI, Intel, NVidia and 
other
brands of GPUs for which a kernel driver exists. All DDX and DIX depend on KMS.
Use of either nomodeset or *.modeset=0 on a kernel command line disables KMS, 
thus
disabling their use.

xf86-video-nouveau is unique among all other X drivers in that it is created and
maintained via reverse engineering, as NVidia refuses to release specifications
required for the FOSS community to conveniently and completely develop device
drivers, quite unlike AMD and Intel, who do most of the actual development for 
the
FOSS DDX drivers amdcpu and intel.

The gist of all the above is that in a fresh installation of Debian that 
excludes
all xserver-xorg-video-, Xorg and Xwayland should just work with all
non-ancient AMD, Intel and NVidia GPUs released before some short period of 
months
or weeks prior to the release of the applicable Debian release, which clearly
includes the GeForce GTX 660.

Once a proprietary NVidia driver installation has been attempted, all bets are
off. Until all traces of an NVidia driver installation are eradicated, a FOSS 
DDX
or DIX cannot be used, though either of the poorly performing non-accelerated
generics, from xserver-xorg-video-fbdev or xserver-xorg-video-vesa, can 
typically
be utilized for the purpose of troubleshooting a DDX or DIX problem.
-- 
Evolution as taught in public schools is, like religion,
based on faith, not based on science.

 Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!

Felix Miata  ***  http://fm.no-ip.com/

Re: Graphical session does not close until unattended-upgrades has applied all updates

[Yvan Masson]

Le 23/04/2021 à 21:07, Yvan Masson a écrit :

Le 23/04/2021 à 20:18, didier gaumet a écrit :


1) Unattended-upgrade at shutdown and after Lightdm is stopped: Perhaps 



a solution would be to :
- stop and disable the unattended-upgrade systemd service
- edit the unattended-upgrades config file to setup the shutdown to false
- write/enable/start your custom systemd service that runs 
/usr/bin/unattented-upgrade only after shutdown has been asked and the 



lightdm systemd service has been stopped


It might work, but I have doubt about my ability to do this properly to 


get reliable updates… I am sure this has already worked for me (in 
previous Debian or Ubuntu, can't remember), certainly with a very 
similar configuration. I have just compared the contents of the package 


from Ubuntu but could not find any notable difference. I will submit a 
bug report so that appropriate people are aware of this.


Done here: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=987431




2) unattended-upgrades notifications while Plymouth is running: I do 
not read Python but trying to understand this python script, I have 
the impression that Plymouth does notify the user that 
unattended-upgrédes is running


You are right, I should have clarify this in my previous answer: I have 


already seen this working (as I said i can't remember if it was in a 
previous Debian or Ubuntu).



Good luck ;-)


Thanks!





OpenPGP_signature
Description: OpenPGP digital signature

Re: dovecot packages

[Christian Schmidt]

Hello,

Michael Grant, 19.04.21:

Just noticed on the dovecot.org site, the latest version of dovecot
seems to be v2.3.14, but on the dovecot package tracker on the debian
site:https://tracker.debian.org/pkg/dovecot, we're at v2.3.13.


 offers packages for Debian, too...

Kind Regards
Christian Schmidt

--
Signature not available.

Re: pci 0000:00:01:0: MSI quirk detected; subordinated MSI disabled ...

[Dan Ritter]

Albretch Mueller wrote: 
> > What you lose this way (besides time) is having apt calculate
> > which supporting packages you need. However,
> 
> > https://packages.debian.org
> 
> > will happily tell you all the dependencies of any given package,
> > and then you can get all the dependencies of each of those, and
> > so on.
> 
>  yes, and "japt" would scrape those pages or even simulate apt to
> figure out all dependencies and then download all the packages you
> need to install locally, off line.

It's a niche need. People who do this at scale keep their own
copies of the entire Debian repository (I do, at work.) People
who do this as one-offs tend not to share your specific needs.


> > You can compile your own kernel with no hardware network
> > drivers.
> 
>  Yes, you can! (tm), but imagine, just imagine, as JOhn Lennon sang,
> that networking would be taken out of the kernel!

... because for almost everyone, it's better in the kernel.


> At times it amazes
> me to discuss with you such issues here. What is the point of using
> security based on IP tables when the active code/js bs is used to even
> probe the keyboard for the passwords you use in case you use the same
> one in your off line and on line machines? They have been using js
> injection even as part of general societal AI bots.

This paragraph made sense up until the word active, and then it
went right off the rails.

You seem to be conflating three different things.

iptables is a firewall mechanism that allows you to reject
unwanted traffic based on properties of IP packets.

If you're running someone else's malicious code, you have lost
your security game already. Don't do that.

"AI" either doesn't exist or has been around for decades,
depending on how you want to define it. Your "they" is
undefined. JavaScript is just a language; there's nothing
special about it that can't be done in Haskell, Forth or
Fortran. Web browsers tend to implement it, but you can also
turn it off altogether -- e.g. I can think of three different
mechanisms that are practical:

Use Firefox with the Disable JavaScript extension.

Use Firefox with ublock Origin and turn off JavaScript
execution.

Use a non-JavaScript equipped browser.

-dsr-

Re: pci 0000:00:01:0: MSI quirk detected; subordinated MSI disabled ...

[Albretch Mueller]

> What you lose this way (besides time) is having apt calculate
> which supporting packages you need. However,

> https://packages.debian.org

> will happily tell you all the dependencies of any given package,
> and then you can get all the dependencies of each of those, and
> so on.

 yes, and "japt" would scrape those pages or even simulate apt to
figure out all dependencies and then download all the packages you
need to install locally, off line.
~
> You can compile your own kernel with no hardware network
> drivers.

 Yes, you can! (tm), but imagine, just imagine, as JOhn Lennon sang,
that networking would be taken out of the kernel! At times it amazes
me to discuss with you such issues here. What is the point of using
security based on IP tables when the active code/js bs is used to even
probe the keyboard for the passwords you use in case you use the same
one in your off line and on line machines? They have been using js
injection even as part of general societal AI bots.
~
> HURD is a microkernel whereby the kernel only provides the most basic
> services and almost all device drivers are implemented in user space.

 https://www.debian.org/ports/hurd/

 https://people.debian.org/~sthibault/failed_packages.txt

 https://www.gnu.org/software/hurd/users-guide/using_gnuhurd.html#Networking

 Thanks and networking doesn't seem to have been taken out of the
kernel in HURD. My only question would be would HURD make it easier
somehow? It would be a project to definitely look into.

 lbrtchx

Re: Graphical session does not close until unattended-upgrades has applied all updates

[Yvan Masson]

Le 23/04/2021 à 20:18, didier gaumet a écrit :


1) Unattended-upgrade at shutdown and after Lightdm is stopped: Perhaps 



a solution would be to :
- stop and disable the unattended-upgrade systemd service
- edit the unattended-upgrades config file to setup the shutdown to false
- write/enable/start your custom systemd service that runs 
/usr/bin/unattented-upgrade only after shutdown has been asked and the 
lightdm systemd service has been stopped


It might work, but I have doubt about my ability to do this properly to 
get reliable updates… I am sure this has already worked for me (in 
previous Debian or Ubuntu, can't remember), certainly with a very 
similar configuration. I have just compared the contents of the package 
from Ubuntu but could not find any notable difference. I will submit a 
bug report so that appropriate people are aware of this.




2) unattended-upgrades notifications while Plymouth is running: I do not 
read Python but trying to understand this python script, I have the 
impression that Plymouth does notify the user that unattended-upgrédes 
is running


You are right, I should have clarify this in my previous answer: I have 
already seen this working (as I said i can't remember if it was in a 
previous Debian or Ubuntu).



Good luck ;-)


Thanks!



OpenPGP_signature
Description: OpenPGP digital signature

Re: installing two versions of a user application

[Stefan Monnier]

> You probably want to look inside the control archives rather than the
> data archives:
>
> https://en.wikipedia.org/wiki/Deb_%28file_format%29
>
> If both control archive files are using a package name of
> "master-pdf-editor", you can try Stefan's suggestion of renaming one
> and seeing if they'll then coinstall.

Exactly: the immediate problem you're encountering is that the package
names themselves (regardless of the files they contain) collide.
This info is the `control` part of the .deb archive.

The way I do it is to use Emacs to open the `foo.deb` file.
This should show you something like:

MMode  Size  Date&time   Filename
- --  --    
  -rw-r--r--   4  2021-03-01 12:23  debian-binary
  -rw-r--r--1128  2021-03-01 12:23  control.tar.xz
  -rw-r--r--  615380  2021-03-01 12:23  data.tar.xz
- --  --    
  616512 3 files

where the `data.tar.xz` part contains the actual files and
`control.tar.xz` is the part you want to modify.
So you select the `control.tar.xz` which itself will look somewhat
similar and will contain a `control` which you can open and in which
you'll find presumably something like:

Package: master-pdf-editor
[...]

followed by other kinds of metainfo about the package.  Just edit that
first line (e.g. by adding a `4` or `5`), then do `C-x C-s` to save this
`control` back into `control.tar.xz`, then kill this buffer with `C-x
k RET`, then save the new `control.tar.xz` back into the actual foo.deb
file with `C-x C-s`.

Then retry to install those debs with:

dpkg --install foo.deb

I recommend you do those modifications on a copy of the file, just so
its easier to recover from mistakes.


Stefan

Re: No networking after resume from suspend

[Dan Ritter]

Richmond wrote: 
> Dan Ritter  writes:
> 
> > Richmond wrote: 
> >> Looks like this bug:
> >> 
> >> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1798921
> >
> > Could be.
> >
> > In which case, this will be solved in the new Stable, most
> > likely out in May or June.
> >
> > You could try a backports kernel before that.
> >
> 
> I tried kernels 5.9 and 5.10 but neither cured it. So perhaps I should report 
> it
> as a bug in debian?

To the kernel packagers group, yup.

-dsr-

Re: Graphical session does not close until unattended-upgrades has applied all updates

[didier gaumet]

1) Unattended-upgrade at shutdown and after Lightdm is stopped: Perhaps 
a solution would be to :

- stop and disable the unattended-upgrade systemd service
- edit the unattended-upgrades config file to setup the shutdown to false
- write/enable/start your custom systemd service that runs 
/usr/bin/unattented-upgrade only after shutdown has been asked and the 
lightdm systemd service has been stopped


2) unattended-upgrades notifications while Plymouth is running: I do not 
read Python but trying to understand this python script, I have the 
impression that Plymouth does notify the user that unattended-upgrédes 
is running


Good luck ;-)

Re: No networking after resume from suspend

[Richmond]

Dan Ritter  writes:

> Richmond wrote: 
>> Looks like this bug:
>> 
>> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1798921
>
> Could be.
>
> In which case, this will be solved in the new Stable, most
> likely out in May or June.
>
> You could try a backports kernel before that.
>

I tried kernels 5.9 and 5.10 but neither cured it. So perhaps I should report it
as a bug in debian?

Re: installing two versions of a user application

[Celejar]

On Fri, 23 Apr 2021 18:48:13 +0100
Morgan Read  wrote:

> On 23/04/2021 6:20 pm, Stefan Monnier wrote:
> > I suggest you look inside those .deb files: they seem to both name the
> > package they install `master-pdf-editor` but dpkg does not allow
> > installing different versions of a given package at the same time.
> > So you'll need to edit those `.deb` to change the package name from
> > `master-pdf-editor` to (say) `master-pdf-editor4` and
> > `master-pdf-editor5`, for example.
> 
> Thanks Stefan, I was wondering about something along those lines, but...
> Inside master-pdf-editor-4.3.89_qt5.x86_64.deb I find:
> data.tar.xz > /./opt/master-pdf-editor-4/masterpdfeditor4
>   /./usr/share/icons/hicolor/128x128/apps/masterpdfeditor4.png
>   /./usr/share/applications/masterpdfeditor4.desktop
> 
> And, inside master-pdf-editor-5.7.08-qt5_included.x86_64.deb I find:
> data.tar.xz > /./opt/master-pdf-editor-5/masterpdfeditor5.sh
>   /./opt/master-pdf-editor-5/masterpdfeditor5.png
>   /./opt/master-pdf-editor-5/masterpdfeditor5
>   /./usr/share/applications/masterpdfeditor5.desktop
>   /./usr/share/icons/hicolor/256x256/apps/masterpdfeditor5.png
> 
> So, everything looks fairly well distinguished - but, I know *nothing*
> about packaging...

You probably want to look inside the control archives rather than the
data archives:

https://en.wikipedia.org/wiki/Deb_%28file_format%29

If both control archive files are using a package name of
"master-pdf-editor", you can try Stefan's suggestion of renaming one
and seeing if they'll then coinstall.

Celejar

Re: installing two versions of a user application

[Celejar]

On Fri, 23 Apr 2021 18:07:02 +0100
Morgan Read  wrote:

> On 23/04/2021 5:40 pm, Celejar wrote:
> > I'll defer to those with a better understanding of the debian packaging
> > system internals, and it may depend on the details of the package, but
> > I'm pretty sure that there's no general way to direct dpkg to install a
> > package to a location other than the one specified by the package. If
> > the original packages were installable to separate directories, then it
> > should be possible to modify the debian packages to work that way as
> > well, but that's well above my pay grade.
> 
> Thanks Celejar, Just to be clear:
> v4 continues to install to /opt/master-pdf-editor-4;
> v5 continues to install to /opt/master-pdf-editor-5.
> But:
> When v4 is installed, /opt/master-pdf-editor-5 is deleted;
> When v5 is installed, /opt/master-pdf-editor-4 is deleted.
> Under the rpm based system, both /opt/master-pdf-editor-5 and
> /opt/master-pdf-editor-4 coexisted.  My memory is fading...  But, iirc,
> it may have been the case that whenever I upgraded v5 I had to reinstall
> v4 as the v5 install removed it.
> 
> The packages that Code Industry provides are clearly designed to get rid
> of v4, obviously because they now want people to buy those features.
> However, installing v4 over v5 never used to remove v5...
> 
> So, I wonder if alien might have done something in the transliteration
> of the v4 rpm to cause the v4 deb to remove other versions when it
> didn't as an rpm - or, is this just the mechanics of the apt system?
> 
> Many thanks.

See Stefan's answer - that's what I meant by modifying the debian
packages.

Celejar

Re: No networking after resume from suspend

[Cindy Sue Causey]

On 4/23/21, Dan Ritter  wrote:
> Richmond wrote:
>> > Let's try from the bottom up?
>> >
>> > ip link show
>> >   will show you the interfaces recognized by the kernel. If this
>> >   works, it might show you an eth0, an en0, or something like a
>> >   enp22s0 device. Let's call it "SAM".
>>
>> Thanks for your reply.
>>
>>  enp2s0:  mtu 1500 qdisc pfifo_fast
>> state DOWN mode DEFAULT group default qlen 1000
>
>
>> >   and make sure the cable is plugged in.
>>
>> :)
>
> Ah, you laugh, but: the NO-CARRIER...UP... state DOWN means that
> the kernel recognizes the device and there's a cable issue.
>
> Go trace the cable, replug in both ends, and if that doesn't
> work, replace the cable. If the far end has a free port, try
> swapping to a different port - this one might be toasted.


Those ethernet to USB port adapter products have been irreplaceable
for me. The only ones that have NEVER worked are the ones that come as
a two-for-one package. They don't seem to work for most other Users,
either. I tested the one I have just a few days ago AGAIN. Zero
interaction.

Maybe I'll try it yet AGAIN today. I'd like to solve that so that it's
one more successful option for mainly poverty folks with computers.
The package I bought was surely under $10. I think I've probably paid
at most maybe $5 or $6 for the single ones.

All of the adapters that come as a single item have worked great. You
just have to track down the new identifier, e.g. one of mine is
"enx00909e9dd1ee". That long value goes wherever one normally types in
eth0, enp1s0, eno1, etc.

Dmesg often reports that value. I grep /var/log/kern* for the word
"renamed" because that's what I always remember first. Except that
then there's having to open kern.log in a text editor occasionally so
yeah, dmesg rocks if one can remember to go that route.

Grep works with dmesg, too. I just tested it. Only one line to read as
output. Very cognitively friendly! :)

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

* runs with birdseed *

Re: bullseye fslint

[Nicholas Geovanis]

On Fri, Apr 23, 2021, 10:30 AM The Wanderer  wrote:

> On 2021-04-23 at 11:27, Greg Wooledge wrote:
>
> > On Fri, Apr 23, 2021 at 07:50:45AM -0400, The Wanderer wrote:
> >
> >> grep -i 'fs|filesystem'" finds rmlint, which looks like it might do
> >> the
> >
> > FYI that needs to be egrep, or grep -E, in order for the | character
> > to work as you intend.
>
> Yeah - sorry, I retyped rather than copying, and didn't realize the
> difference it made until I'd already hit Send.
>
> It always seems bizarre to me that that would be one of the characters
> that needs escaping to have a special function in standard regex, and
> consequently I always have trouble remembering it...
>

If it helps, they're both meta-characters. Logical notation embedded in a
command stream. So both need to "escape" into logic from command-line
reality :-)

   The Wanderer
>
> The reasonable man adapts himself to the world; the unreasonable one
> persists in trying to adapt the world to himself. Therefore all
> progress depends on the unreasonable man. -- George Bernard Shaw
>
>

Re: installing two versions of a user application

[Stefan Monnier]

> $ sudo apt install ./master-pdf-editor-4.3.89_qt5.x86_64.deb
> ./master-pdf-editor-5.7.08-qt5_included.x86_64.deb
> Reading package lists... Done
> Building dependency tree
> Reading state information... Done
> Note, selecting 'master-pdf-editor' instead of
> './master-pdf-editor-4.3.89_qt5.x86_64.deb'
> Note, selecting 'master-pdf-editor' instead of
> './master-pdf-editor-5.7.08-qt5_included.x86_64.deb'
> master-pdf-editor is already the newest version (4.3.89-2).
> The following packages will be upgraded:
>   master-pdf-editor

I suggest you look inside those .deb files: they seem to both name the
package they install `master-pdf-editor` but dpkg does not allow
installing different versions of a given package at the same time.
So you'll need to edit those `.deb` to change the package name from
`master-pdf-editor` to (say) `master-pdf-editor4` and
`master-pdf-editor5`, for example.


Stefan "who does that kind of thing in order to install emacs25,
emacs26, and emacs27 at the same time"

Re: No networking after resume from suspend

[Cindy Sue Causey]

On 4/23/21, David Wright  wrote:
> On Fri 23 Apr 2021 at 10:49:00 (+0100), Richmond wrote:
>> Cindy Sue Causey  writes:
>>
>> > Questions where answers might help come to mind. Primarily, has this
>> > always occurred, or did it just start up in the last couple days?
>>
>> It has occured since installing debian (10). Prior to that I was using
>> opensuse.
>>
>> > I'm on a new old secondhand laptop today. Just started using it
>> > yesterday. It's doing similar, but I can trigger it back on via
>> > "wicd-curses". A quick CTRL+R (refresh) for no particular reason and
>> > then CTRL+C (connect), and I'm back up and running.
>>
>> I don't think I am using Wicked, if that is the equivalent of the
>> opensuse Wicked. I think I am using network manager.
>
> JFTR there's no connection (pun, sorry) between Wicd¹ and Wicked.²
> Confusion is compounded by their identical pronunciation.
>
> Wicked would claim to be a superset of NetworkManager. Wicd is a
> simple connection manager, handling just one connection at a time.
>
> In any case, wicd's days are numbered unless and until it moves
> forward from Python2. There's a Python3 version in experimental,
> but the hiatus might not be good for sustaining its popularity.


Experimental is the version I'm using. That was an empowering moment
there. It helped me finally understand all the chatter about backports
and the irreplaceable value to Users looking for missing programs that
didn't make it in time to be included in any given release.


> ¹ acronym, Wireless Interface Connection Daemon
>
> ² "No REST for the wicked", motivational joke after SUSE rejected
>   using REST, a REpresentational State Transfer interface.


Thank you! I always wondered but never pursued what the wicd acronym
was. I assumed it was an acronym, anyway, because of that "d" that
appears in other packages, too.

#1 I'm so sorry, I made a major mistake last night. To operate
wicd-curses, the commands are SHIFT+R for refresh and SHIFT+C to
connect. There's no CTRL used there. That was the first thought I had
this morning after waking up.

The tl;dr for the rest of this is that systemctl worked for me this
morning, but I'm using wicd instead of e.g. network-manager. They're
both still available on Debian because there are differences in how
they function. They don't appear to be duplicate overkill by both
being installable. My memory recall is that they don't play nice
together.

Ok, so I tried those various systemctl lines. When I actually read
them, I realized this isn't an apples to apples comparison. That
"network-manager" was the sign.

In addition to that, I'm on a wired connection in case this was about
wireless. For the moment, I can only assume wicd might handle wireless
connections similarly.

With respect to those terminal commands, I first tried wicd-curses
instead of network-manager. That didn't work because it's an
interface.

So I tried "systemctl restart wicd" but couldn't tell that anything
was happening beyond that it didn't throw any errors. I tried
"systemctl stop wicd", and it understandably killed my open
wicd-curses interface in the terminal.

Next, "systemctl start wicd" was entered. I reopened wicd-curses...
and it was already connected online without any further intervention.
It was cool to see that interaction, am just waiting for my brain to
have the ah-ha moment about exactly what happened.

In the meantime, that's my "learn something new" today. One my
earliest "thank you" posts about Debian was about the fact that there
are frequently at least two ways to accomplish many tasks. That kept
coming to mind over the last month of horrific hardware failures.

Several of those alternates were used repeatedly the last few weeks.
In particular, the right click on the desktop to access the
Applications menu was invaluable. I was working on a badly damaged
screen that blocked access to the usual method of clicking the word
"Applications" on the top left in XFCE4.

N.B. Apparently the dogs or I stepped on my Dell Duo when it was on
the floor beside me here. The leaking damage on the screen looked just
like a giant footprint, lol.

AND I just now remembered that I could have moved the desktop panel
that holds that Application menu. Hopefully there will never be a next
time where that needs to occur.

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

* runs with birdseed *

Re: Nvidia graphic card and newer kernels (open source driver)

[Viktor Vogel]

Just a data point - I'd been using nvidia cards for years.

Onmy current workstation, a Dell XPS 8930, the nvidia binary blob worked 
well, unless and until I booted a kernel other than the straight vanilla 
from the distro. After upgrading to the low latency kernel, it was 
goodbye GUI. So I reverted to the nouveau drivers, which worked fairly 
well for normal activities, and even games like openarena, but crashed 
immediately if I tried launch the game known as Beyond a Steel Sky.


So, caught between a rock and a hard place, I removed my nvidia geforce 
1660 TI and replaced it with a radeon rx 560, since the default amdgpu 
drivers were more capable and performant than the nouveau, and less 
restrictive than the nvidia binary blob.


Personally, I'm not really into nvidia video these days. The default 
intel video is fine for normal desktop activities, and amd is my goto 
for extra performance.


Viktor

On 4/23/21 9:31 AM, Malte Marwedel wrote:
Has anyone experience with Nvidia cards without closed source drivers 
and newer kernels?
Recently my AMD card died, so I got a fast replacement from a friend - 
a GeForce GTX 660.
With the kernel 5.5.9, the PC stops booting after GRUB. The kernel 
tries to set the resolution of the console and stops. X is not even 
attempting to start.

Therefore I checked different kernels:
5.11.16: Not working
5.5.9: Not working
5.4.114:  Working, X is starting and I even have some 3D acceleration 
(kicad 3D view, Extreme tux racer, however Civilization 6 is crashing 
the system).

5.2.9: Working.

With the working kernels, I however see the following error:
Nvidiafb: cannot requests PCI regions
But the system continues to work. So I tried to compile 5.11.16 
without this module - no change compared to 5.5.9.


Any Ideas? I don't want to install the closed source module, I had 
enough pain with fglrx ~10 years ago.


--
"It's always rewarding to talk to a clever man" - Raskolnikov, "Crime and 
Punishment"

Re: installing two versions of a user application

[Celejar]

On Fri, 23 Apr 2021 17:22:48 +0100
Morgan Read  wrote:

> Hello Folks,
> 
> Some little time ago I made a jump from another distro into Debian.
> 
> When I made that jump I took with me a few proprietary applications that
> I make a lot of use of.  One in particular is causing me a bit of a
> headache - Master PDF Editor by Code Industry:
> https://code-industry.net/free-pdf-editor/
> 
> Master PDF Editor is now well into version 5, but when it transitioned
> from v4 to v5 it left a lot of its functionality (in the no cost
> version) behind.  So, I'd always kept a copy of the old v4 rpm to hand
> and installed alongside v5 in my old distro.  Version 4 has long since
> become unavailable, but I was able to make good use of alien to magic my
> v4 rpm into a v4 deb.
> 
> BUT, now I can not install both v4 and v5 alongside one another.  When I
> install v4, v5 is removed; when I install v5, v4 is removed.  This
> wasn't a problem under the old distro.  Each version is installed into
> its own directory, either: /opt/master-pdf-editor-4 or
> /opt/master-pdf-editor-5.
> 
> I tried to force it with the following, but...
> $ sudo apt install ./master-pdf-editor-4.3.89_qt5.x86_64.deb
> ./master-pdf-editor-5.7.08-qt5_included.x86_64.deb
> Reading package lists... Done
> Building dependency tree
> Reading state information... Done
> Note, selecting 'master-pdf-editor' instead of
> './master-pdf-editor-4.3.89_qt5.x86_64.deb'
> Note, selecting 'master-pdf-editor' instead of
> './master-pdf-editor-5.7.08-qt5_included.x86_64.deb'
> master-pdf-editor is already the newest version (4.3.89-2).
> The following packages will be upgraded:
>   master-pdf-editor
> 1 upgraded, 0 newly installed, 0 to remove and 1 not upgraded.
> Need to get 0 B/28.4 MB of archives.
> After this operation, 64.8 MB of additional disk space will be used.
> Do you want to continue? [Y/n] n
> Abort.
> 
> Is there any method to allow v4 and v5 to be both installed together?

I'll defer to those with a better understanding of the debian packaging
system internals, and it may depend on the details of the package, but
I'm pretty sure that there's no general way to direct dpkg to install a
package to a location other than the one specified by the package. If
the original packages were installable to separate directories, then it
should be possible to modify the debian packages to work that way as
well, but that's well above my pay grade.

This does seem like a pretty good use case for containerization or
virtualization. LXC (or some sort of chroot solution) in particular
might be a good fit here - you get a separate filesystem, without the
overhead of a full VM, which shouldn't be necessary in this case.

Celejar

Nvidia graphic card and newer kernels (open source driver)

[Malte Marwedel]

Has anyone experience with Nvidia cards without closed source drivers 
and newer kernels?
Recently my AMD card died, so I got a fast replacement from a friend - a 
GeForce GTX 660.
With the kernel 5.5.9, the PC stops booting after GRUB. The kernel tries 
to set the resolution of the console and stops. X is not even attempting 
to start.

Therefore I checked different kernels:
5.11.16: Not working
5.5.9: Not working
5.4.114:  Working, X is starting and I even have some 3D acceleration 
(kicad 3D view, Extreme tux racer, however Civilization 6 is crashing 
the system).

5.2.9: Working.

With the working kernels, I however see the following error:
Nvidiafb: cannot requests PCI regions
But the system continues to work. So I tried to compile 5.11.16 without 
this module - no change compared to 5.5.9.


Any Ideas? I don't want to install the closed source module, I had 
enough pain with fglrx ~10 years ago.

installing two versions of a user application

[Morgan Read]

Hello Folks,

Some little time ago I made a jump from another distro into Debian.

When I made that jump I took with me a few proprietary applications that
I make a lot of use of.  One in particular is causing me a bit of a
headache - Master PDF Editor by Code Industry:
https://code-industry.net/free-pdf-editor/

Master PDF Editor is now well into version 5, but when it transitioned
from v4 to v5 it left a lot of its functionality (in the no cost
version) behind.  So, I'd always kept a copy of the old v4 rpm to hand
and installed alongside v5 in my old distro.  Version 4 has long since
become unavailable, but I was able to make good use of alien to magic my
v4 rpm into a v4 deb.

BUT, now I can not install both v4 and v5 alongside one another.  When I
install v4, v5 is removed; when I install v5, v4 is removed.  This
wasn't a problem under the old distro.  Each version is installed into
its own directory, either: /opt/master-pdf-editor-4 or
/opt/master-pdf-editor-5.

I tried to force it with the following, but...
$ sudo apt install ./master-pdf-editor-4.3.89_qt5.x86_64.deb
./master-pdf-editor-5.7.08-qt5_included.x86_64.deb
Reading package lists... Done
Building dependency tree
Reading state information... Done
Note, selecting 'master-pdf-editor' instead of
'./master-pdf-editor-4.3.89_qt5.x86_64.deb'
Note, selecting 'master-pdf-editor' instead of
'./master-pdf-editor-5.7.08-qt5_included.x86_64.deb'
master-pdf-editor is already the newest version (4.3.89-2).
The following packages will be upgraded:
  master-pdf-editor
1 upgraded, 0 newly installed, 0 to remove and 1 not upgraded.
Need to get 0 B/28.4 MB of archives.
After this operation, 64.8 MB of additional disk space will be used.
Do you want to continue? [Y/n] n
Abort.

Is there any method to allow v4 and v5 to be both installed together?

Thanks
Morgan
<>

signature.asc
Description: OpenPGP digital signature

installing two versions of a user application

[Morgan Read]

Hello Folks,

Some little time ago I made a jump from another distro into Debian.

When I made that jump I took with me a few proprietary applications that
I make a lot of use of.  One in particular is causing me a bit of a
headache - Master PDF Editor by Code Industry:
https://code-industry.net/free-pdf-editor/

Master PDF Editor is now well into version 5, but when it transitioned
from v4 to v5 it left a lot of its functionality (in the no cost
version) behind.  So, I'd always kept a copy of the old v4 rpm to hand
and installed alongside v5 in my old distro.  Version 4 has long since
become unavailable, but I was able to make good use of alien to magic my
v4 rpm into a v4 deb.

BUT, now I can not install both v4 and v5 alongside one another.  When I
install v4, v5 is removed; when I install v5, v4 is removed.  This
wasn't a problem under the old distro.  Each version is installed into
its own directory, either: /opt/master-pdf-editor-4 or
/opt/master-pdf-editor-5.

I tried to force it with the following, but...
$ sudo apt install ./master-pdf-editor-4.3.89_qt5.x86_64.deb
./master-pdf-editor-5.7.08-qt5_included.x86_64.deb
Reading package lists... Done
Building dependency tree
Reading state information... Done
Note, selecting 'master-pdf-editor' instead of
'./master-pdf-editor-4.3.89_qt5.x86_64.deb'
Note, selecting 'master-pdf-editor' instead of
'./master-pdf-editor-5.7.08-qt5_included.x86_64.deb'
master-pdf-editor is already the newest version (4.3.89-2).
The following packages will be upgraded:
  master-pdf-editor
1 upgraded, 0 newly installed, 0 to remove and 1 not upgraded.
Need to get 0 B/28.4 MB of archives.
After this operation, 64.8 MB of additional disk space will be used.
Do you want to continue? [Y/n] n
Abort.

Is there any method to allow v4 and v5 to be both installed together?

Thanks
Morgan
<>

signature.asc
Description: OpenPGP digital signature

Re: No networking after resume from suspend

[David Wright]

On Fri 23 Apr 2021 at 10:49:00 (+0100), Richmond wrote:
> Cindy Sue Causey  writes:
> 
> > Questions where answers might help come to mind. Primarily, has this
> > always occurred, or did it just start up in the last couple days?
> 
> It has occured since installing debian (10). Prior to that I was using
> opensuse.
> 
> > I'm on a new old secondhand laptop today. Just started using it
> > yesterday. It's doing similar, but I can trigger it back on via
> > "wicd-curses". A quick CTRL+R (refresh) for no particular reason and
> > then CTRL+C (connect), and I'm back up and running.
> 
> I don't think I am using Wicked, if that is the equivalent of the
> opensuse Wicked. I think I am using network manager.

JFTR there's no connection (pun, sorry) between Wicd¹ and Wicked.²
Confusion is compounded by their identical pronunciation.

Wicked would claim to be a superset of NetworkManager. Wicd is a
simple connection manager, handling just one connection at a time.

In any case, wicd's days are numbered unless and until it moves
forward from Python2. There's a Python3 version in experimental,
but the hiatus might not be good for sustaining its popularity.

¹ acronym, Wireless Interface Connection Daemon

² "No REST for the wicked", motivational joke after SUSE rejected
  using REST, a REpresentational State Transfer interface.

Cheers,
David.

Re: bullseye fslint

[Greg Wooledge]

On Fri, Apr 23, 2021 at 11:29:59AM -0400, The Wanderer wrote:
> On 2021-04-23 at 11:27, Greg Wooledge wrote:
> 
> > On Fri, Apr 23, 2021 at 07:50:45AM -0400, The Wanderer wrote:
> > 
> >> grep -i 'fs|filesystem'" finds rmlint, which looks like it might do
> >> the
> > 
> > FYI that needs to be egrep, or grep -E, in order for the | character 
> > to work as you intend.
> 
> Yeah - sorry, I retyped rather than copying, and didn't realize the
> difference it made until I'd already hit Send.
> 
> It always seems bizarre to me that that would be one of the characters
> that needs escaping to have a special function in standard regex, and
> consequently I always have trouble remembering it...

It's all historical.  There's no rhyme or reason to it, except that
various tools worked this way starting in the 1970s.  Then some merging
and standardization was done, with the result that there are two standard
regex flavors in POSIX -- Basic and Extended.

grep and sed use BRE (Basic Regular Expressions) by default, and in BRE
you don't have | ? + operators, and you use \{ \} for expressing counts.

grep -E and awk use ERE (Extended Regular Expressions), which gives you
the | ? + operators, and you use { } for counts and ( ) for grouping.

It gets worse the deeper you dive into it, with PCRE and GNU extensions
flying around, a proposal to add -E to sed, etc.

Re: bullseye fslint

[The Wanderer]

On 2021-04-23 at 11:27, Greg Wooledge wrote:

> On Fri, Apr 23, 2021 at 07:50:45AM -0400, The Wanderer wrote:
> 
>> grep -i 'fs|filesystem'" finds rmlint, which looks like it might do
>> the
> 
> FYI that needs to be egrep, or grep -E, in order for the | character 
> to work as you intend.

Yeah - sorry, I retyped rather than copying, and didn't realize the
difference it made until I'd already hit Send.

It always seems bizarre to me that that would be one of the characters
that needs escaping to have a special function in standard regex, and
consequently I always have trouble remembering it...

-- 
   The Wanderer

The reasonable man adapts himself to the world; the unreasonable one
persists in trying to adapt the world to himself. Therefore all
progress depends on the unreasonable man. -- George Bernard Shaw



signature.asc
Description: OpenPGP digital signature

Re: bullseye fslint

[Greg Wooledge]

On Fri, Apr 23, 2021 at 07:50:45AM -0400, The Wanderer wrote:
> grep -i 'fs|filesystem'" finds rmlint, which looks like it might do the

FYI that needs to be egrep, or grep -E, in order for the | character
to work as you intend.

Re: No networking after resume from suspend

[Dan Ritter]

Richmond wrote: 
> Looks like this bug:
> 
> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1798921

Could be.

In which case, this will be solved in the new Stable, most
likely out in May or June.

You could try a backports kernel before that.

-dsr-

Re: No networking after resume from suspend

[Richmond]

Looks like this bug:

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1798921

Re: No networking after resume from suspend

[Dan Ritter]

Richmond wrote: 
> > Let's try from the bottom up?
> >
> > ip link show
> >   will show you the interfaces recognized by the kernel. If this
> >   works, it might show you an eth0, an en0, or something like a 
> >   enp22s0 device. Let's call it "SAM".
> 
> Thanks for your reply.
> 
>  enp2s0:  mtu 1500 qdisc pfifo_fast state 
> DOWN mode DEFAULT group default qlen 1000


> >   and make sure the cable is plugged in.
> 
> :)

Ah, you laugh, but: the NO-CARRIER...UP... state DOWN means that
the kernel recognizes the device and there's a cable issue.

Go trace the cable, replug in both ends, and if that doesn't
work, replace the cable. If the far end has a free port, try
swapping to a different port - this one might be toasted.

-dsr-

Re: Graphical session does not close until unattended-upgrades has applied all updates

[Yvan Masson]

Le 23/04/2021 à 12:14, didier gaumet a écrit :

Le 23/04/2021 à 11:59, didier gaumet a écrit :

I should check what I wrote before posting :-(

[...]

Disclaimer: I have never experimented myself whait I suggest


"what"

[...]
As for Plymouth notifying about ugrades during shutdown, if it is not 
already the case (when unattended-upgrades the 


A bit is missing there (cut/past and absent-minded poster): ..." systemd 
service is started after the LightDM systemd service is stopped), it is 



perhaps possible to modify the "...

unattended-upgrades.service systemd service to script Plymouth, or 
more simply, setup Plymouth to be more informative

http://blog.fpmurphy.com/2009/09/project-plymouth.html


Thanks Didier for your answer, it looked promising. Unfortunately I 
suppose it is not the way to go: unattended-upgrades.service starts 
`/usr/share/unattended-upgrades/unattended-upgrade-shutdown` at boot, 
which waits for "a signal". (I suppose from reading the code that it 
listen on dbus something related to systemd-logind, but my skills 
prevent me from going further). Anyway, in my case it starts properly, 
but it just prevents the session/LightDM from stopping. If no other 
idea, I suppose I should submit a bug report?


Regards,
Yvan



OpenPGP_signature
Description: OpenPGP digital signature

Re: bullseye fslint

[Peter Ehlert]

On 4/23/21 4:50 AM, The Wanderer wrote:

On 2021-04-23 at 07:29, Peter Ehlert wrote:


fslint is gone. what alternatives are there?

I've never used such a tool myself, but a quick "apt-cache search lint |
grep -i 'fs|filesystem'" finds rmlint, which looks like it might do the
job...


I want a GUI in the debian repos

...and Recommends: rmlint-gui, so while I can't speak to its quality,
apparently that exists as well.

thanks.

the GUI launcher in Mate  applications > system tools > Shredder 
Duplicate Finder


the GUI is way different that fslint but perhaps it will function as desired


(At a glance, just from the 'apt-cache show fslint' output, I'd guess
that the reason fslint is gone is that it's Python-2-only and Python 2
is going away in bullseye.)

Re: pci 0000:00:01:0: MSI quirk detected; subordinated MSI disabled ...

[Darac Marjal]

On 22/04/2021 20:53, Albretch Mueller wrote:
[cut]

>  I would also love to see networking taken out of the Linux kernel,
> but this is an entirely different, hellishly "political" issue.
It's not quite the same thing, but you might be able to get what you
want with the Debian HURD port .
HURD is a microkernel whereby the kernel only provides the most basic
services and almost all device drivers are implemented in user space.
>
>  lbrtchx
>



OpenPGP_signature
Description: OpenPGP digital signature

Re: bullseye fslint

[The Wanderer]

On 2021-04-23 at 07:29, Peter Ehlert wrote:

> fslint is gone. what alternatives are there?

I've never used such a tool myself, but a quick "apt-cache search lint |
grep -i 'fs|filesystem'" finds rmlint, which looks like it might do the
job...

> I want a GUI in the debian repos

...and Recommends: rmlint-gui, so while I can't speak to its quality,
apparently that exists as well.

(At a glance, just from the 'apt-cache show fslint' output, I'd guess
that the reason fslint is gone is that it's Python-2-only and Python 2
is going away in bullseye.)

-- 
   The Wanderer

The reasonable man adapts himself to the world; the unreasonable one
persists in trying to adapt the world to himself. Therefore all
progress depends on the unreasonable man. -- George Bernard Shaw



signature.asc
Description: OpenPGP digital signature

bullseye fslint

[Peter Ehlert]

fslint is gone. what alternatives are there?
I want a GUI in the debian repos

Re: No networking after resume from suspend

[Andrei POPESCU]

On Vi, 23 apr 21, 10:46:21, Richmond wrote:
> Dan Ritter  writes:
> >
> >   If any part of SAM reads "down", do this:
> >
> > sudo ip link set up SAM
> 
> ip link set up enp2s0
> 
> No output here but the link is still down.

This suggests to me the problem is at a lower level, i.e. the kernel 
module (driver).

Try unloading / re-loading it ('modprobe -r' and 'modprobe').

You should be able to find the module in use with 'lspci -k', possibly 
with '-nn' as well to help identify the correct device[1], or by 
examining the kernel log (dmesg).

[1] can't test this as the hardware in use here doesn't have PCI buses.

Kind regards,
Andrei
-- 
http://wiki.debian.org/FAQsFromDebianUser


signature.asc
Description: PGP signature

Re: Graphical session does not close until unattended-upgrades has applied all updates

[didier gaumet]

Le 23/04/2021 à 11:59, didier gaumet a écrit :

I should check what I wrote before posting :-(

[...]

Disclaimer: I have never experimented myself whait I suggest


"what"

[...]
As for Plymouth notifying about ugrades during shutdown, if it is not 
already the case (when unattended-upgrades the 


A bit is missing there (cut/past and absent-minded poster): ..." systemd 
service is started after the LightDM systemd service is stopped), it is 
perhaps possible to modify the "...


unattended-upgrades.service systemd service to script Plymouth, or more 
simply, setup Plymouth to be more informative

http://blog.fpmurphy.com/2009/09/project-plymouth.html

Re: No networking after resume from suspend

[Richmond]

Dan Ritter  writes:

> Richmond wrote: 
>> When I resume from suspend there is no networking. I cannot find a way
>> to restart it. I tried these various commands.
>> 
>> systemctl restart network
>> /etc/init.d/networking restart
>> systemctl reset-failed 
>> systemctl restart networking.service 
>> systemctl restart network-online.target 
>> systemctl restart network-manager.service 
>> systemctl start network-manager.service 
>> systemctl stop network-manager.service 
>> 
>> lshw says:
>> 
>>  description: Ethernet interface
>>  product: 88E8071 PCI-E Gigabit Ethernet Controller
>>  vendor: Marvell Technology Group Ltd.
>
> Let's try from the bottom up?
>
> ip link show
>   will show you the interfaces recognized by the kernel. If this
>   works, it might show you an eth0, an en0, or something like a 
>   enp22s0 device. Let's call it "SAM".

Thanks for your reply.

 enp2s0:  mtu 1500 qdisc pfifo_fast state 
DOWN mode DEFAULT group default qlen 1000
 
I omited the second line which contained what looks like a mac address.

>
>   If any part of SAM reads "down", do this:
>
> sudo ip link set up SAM

ip link set up enp2s0

No output here but the link is still down.



>
>   and make sure the cable is plugged in.

:)

>
> Now let's set an address.
>
> ip addr show SAM
>   that might show you an IPv4 address. If not, or if you don't
>   think it looks right for your network, you can set one by
>   hand:

No address here, it repeats that the link is down.

>
> sudo ip addr set 192.168.0.43/24 dev SAM
>
>   or you can ask DHCP for one:
>
> sudo dhclient SAM
>

This paused for a long time but didn't set an address.

So I will try to set one manually as above...

It didn't like 'set', so I used 'add'

The address it is supposed to be is 192.168.1.43, you came close to
guessing it!

ip addr add 192.168.1.43/24 dev enp2s0

This produced a message in the GUI saying a connection was established
to enp2s0 but the link is still down.

> Finally, try pinging your gateway router.
>
> Let us know how far you get with this, and exactly what the
> error messages are.
>

ping 192.168.1.1
PING 192.168.1.1 (192.168.1.1) 56(84) bytes of data.
>From 192.168.1.43 icmp_seq=1 Destination Host Unreachable
... etc

I tried this before the other commands too, and got this:

ping 192.168.1.1
connect: Network is unreachable

Re: No networking after resume from suspend

[Richmond]

Thanks for your reply.

Cindy Sue Causey  writes:

> Questions where answers might help come to mind. Primarily, has this
> always occurred, or did it just start up in the last couple days?

It has occured since installing debian (10). Prior to that I was using
opensuse.

>
> I'm on a new old secondhand laptop today. Just started using it
> yesterday. It's doing similar, but I can trigger it back on via
> "wicd-curses". A quick CTRL+R (refresh) for no particular reason and
> then CTRL+C (connect), and I'm back up and running.

I don't think I am using Wicked, if that is the equivalent of the
opensuse Wicked. I think I am using network manager.

There is nothing in

/etc/network/interfaces.d/

I haven't messed around with things I just installed debian 10 with Mate
and it mostly works.

Re: debian-user-digest Digest V2021 #410

[Natalie Grobler]

Unsubscribe

On Tue, Apr 20, 2021 at 1:50 PM 
wrote:

> Content-Type: text/plain
>
> debian-user-digest Digest   Volume 2021 :
> Issue 410
>
> Today's Topics:
>   Re: Postfix configuration on Bullsey  [ Darac Marjal <
> mailingl...@darac.org ]
>   Buster Apache php website in other l  [ john doe 
> ]
>   Re: Strange emacs behavior after upg  [ Rainer Dorsch  ]
>   Re: Strange emacs behavior after upg  [ Rainer Dorsch  ]
>   Re: Strange emacs behavior after upg  [  ]
>   Re: Buster Apache php website in oth  [ IL Ka 
> ]
>   Apology; misleading statement about   [ rhkra...@gmail.com ]
>   Re: Strange emacs behavior after upg  [ Greg Wooledge 
> ]
> Date: Tue, 20 Apr 2021 08:51:29 +0100
> From: Darac Marjal 
> To: debian-user@lists.debian.org
> Subject: Re: Postfix configuration on Bullseye
> Message-ID: 
> Content-Type: multipart/signed; micalg=pgp-sha256;
>  protocol="application/pgp-signature";
>  boundary="cXiLyYr1h7hktnun1qRKqfgTA5kxJnbj7"
>
> This is an OpenPGP/MIME signed message (RFC 4880 and 3156)
> --cXiLyYr1h7hktnun1qRKqfgTA5kxJnbj7
> Content-Type: multipart/mixed;
> boundary="vaHTE3qsZ9Kr7kII2tmvpoZf3QCu1zFL7";
>  protected-headers="v1"
> From: Darac Marjal 
> To: debian-user@lists.debian.org
> Message-ID: 
> Subject: Re: Postfix configuration on Bullseye
> References: <20210419170855.3e7a4b3d@hawk.localdomain>
> In-Reply-To: <20210419170855.3e7a4b3d@hawk.localdomain>
>
> --vaHTE3qsZ9Kr7kII2tmvpoZf3QCu1zFL7
> Content-Type: text/plain; charset=utf-8
> Content-Transfer-Encoding: quoted-printable
> Content-Language: en-GB
>
>
> On 20/04/2021 00:08, Charles Curley wrote:
> > On installing on Bullseye, I usually install postfix, then configure it=
>
> > with "dpkg-reconfigure postfix".
> >
> > I use postfix here only for logwatch and other system emails, so the
> > setup isn't concerned with the Internet at large.
> >
> > The default list of systems to accept mail for doesn't look right to me=
> :
> >
> > grissom.localdomain, grissom.localdomain, localhost.localdomain, , loca=
> lhost
> >
> > * Why is the fully qualified host name in there twice, but not the
> >   hostname alone ("grissom")? (localdomain is my local TLD on a private=
>
> >   network.)
> >
> > * What with the two commas toward the end?
> >
> > Shouldn't that be
> >
> > grissom.localdomain, grissom, localhost.localdomain, localhost
>
> This looks to come from the debian/postfix.config file, and is thus part
> of the Debian packaging of postfix, rather than an upstream thing. In
> that file, at line 228, we see:=C2=A0
>
> if ($mailertype eq "Internet Site") { if ($mailname eq $hostname) {
> $destinations =3D join ", ",("\$myhostname", $mailname, "localhost." .
> $domain, ", localhost"); } else { $destinations =3D join ",
> ",("\$myhostname", $mailname, $hostname, "localhost." . $domain . ",
> localhost"); } } else { # don't accept mail for $mailname by default if
> we have a relayhost or local only mail, # unless the mailname bears no
> resemblance to $myorigin. $destinations =3D join ", ",("\$myhostname",
> $hostname, "localhost." . $domain . ", localhost" ); unless ( $hostname
> =3D~ m/(^|[\.])$mailname$/ ) { $destinations =3D $mailname . ", " .
> $destinations; } }
>
> [ Taken from
> https://sources.debian.org/src/postfix/3.5.6-1/debian/postfix.config/#L22=
> 8,
> which might be easier to read if that wraps ]
>
> This is perl, so the join() function takes a string and an array and
> delimits the array with the string. So, if we take the first one as an
> example, the literal string "$myhostname" is followed by a comma-space,
> then the value in the "mailname" variable, then the literal string
> "localhost." with the "domain" variable appended, then another
> comma-space. Finally, the last element to be added to the list is ",
> localhost". I don't know why this was written this way, but it means
> that in every case, the "destinations" variable will end with ", ,
> localhost"
>
> Sadly, the earliest revision I can find of this file on salsa.debian.org
> (https://salsa.debian.org/postfix-team/postfix-dev/-/commit/a0577ca96dda9=
> c4e5e5bc9dd0c5b7cfc545c5804#ac03215119d5f2efaeb830653c7f84124ceed640_0_19=
> 
> 2)
> already has the ", localhost" code in it, so I can't say why it was
> written like that.
>
> On the upside, though, this is an allowlist of domains postfix will
> accept mail for. If there are duplicates, it shouldn't REALLY make much
> difference. It's a nice to fix (just because, if you can't explain why
> the code is doing something weird, you can't adequately say whether it's
> a problem or not).
>
>
>
>
>
> --vaHTE3qsZ9Kr7kII2tmvpoZf3QCu1zFL7--
>
> --cXiLyYr1h7hktnun1qRKqfgTA5kxJnbj7
> Content-Type: application/pgp-signature; name="OpenPGP_signature.asc"
> Content-Description: OpenPGP digital signature
> Content-Disposition: attachment; filename="O

Re: Graphical session does not close until unattended-upgrades has applied all updates

[didier gaumet]

Le 23/04/2021 à 09:48, Yvan Masson a écrit :
[...]
I would expect the session to be completely closed, LightDM stopped, and 
the console or Plymouth displaying a message indicating the ongoing 
updates.

[...]

Hello,

Disclaimer: I have never experimented myself whait I suggest

Perhaps adapting the unattended-upgrades.service systemd service by 
declaring that the lightdm service is to be stopped before this service 
is started would do.


As for Plymouth notifying about ugrades during shutdown, if it is not 
already the case (when unattended-upgrades the 
unattended-upgrades.service systemd service to script Plymouth, or more 
simply, setup Plymouth to be more informative

http://blog.fpmurphy.com/2009/09/project-plymouth.html

Re: Any Debian BPMN editor (or a converter)?

[Michael-John Turner]

On Wed, Apr 21, 2021 at 04:24:33PM -0300, Markos wrote:

Any suggetion about an alternative of editor or a converter that converts a
file in the format, for example svg (or any other), to bpmn?


I can't comment specifically on how well it works importing SVGs as I've
not used that functionality, but diagrams.net[1] certainly supports both
SVG import and editing BPMN diagrams. I've used it for the latter with good
success. It's open source and available both hosted and standalone.

[1] https://www.diagrams.net/

Cheers, MJ
--
Michael-John Turner * m...@mjturner.net * http://mjturner.net/ 

Graphical session does not close until unattended-upgrades has applied all updates

[Yvan Masson]

Hi list,

I am preparing some Debian 11 desktops for a school (for when it will be 
the new stable). The setup is very simple: no root account, one 
partition, tasks desktop/Cinnamon/SSH. Unattended-upgrades is configured 
to install updates on shutdown (see 1): upgrading works properly, but is 
very disturbing for the users:


When the user chooses to shutdown or reboot the computer from his 
Cinnamon session, the session does not close until all updates are 
applied. While waiting:

- icons on the desktop disappear
- the usual menu that allows choosing between 
suspend/hibernate/reboot/cancel/shutdown won't appear again (see 2)

- it is still possible to start applications

When a user session has been opened, then closed, and the user clicks on 
shutdown or reboot from LightDM, the behavior is similar: LightDM does 
not stop. It is even possible to log in again, while unattended-upgrades 
is applying updates, but when updates are applied the computer 
shutdowns/reboots as requested originally from LightDM.


I would expect the session to be completely closed, LightDM stopped, and 
the console or Plymouth displaying a message indicating the ongoing 
updates. I am almost sure this works like this on Ubuntu, with the same 
setup from me. Where should I look to solve this? Should I report a bug, 
and on which package?


Regards,
Yvan


1. Here are the files I create to configure unattended-upgrades:
$ cat /etc/apt/apt.conf.d/21periodic
  APT::Periodic::Unattended-Upgrade "1";
  APT::Periodic::CleanInterval "1";

$ cat /etc/apt/apt.conf.d/51unattended-upgrades-local
Unattended-Upgrade::Origins-Pattern {
  "origin=Debian,codename=bullseye,label=Debian";
  "origin=Debian,codename=bullseye,label=Debian-Security";
  "origin=Debian,codename=bullseye-updates,label=Debian";
};
Unattended-Upgrade::Mail "root";
nattended-Upgrade::InstallOnShutdown "true";
Unattended-Upgrade::Remove-Unused-Kernel-Packages "false";

$ cat /etc/systemd/logind.conf.d/unattended-upgrades.conf
[Login]
# 30 minutes delay
InhibitDelayMaxSec=1800


2. This matches the following error in ~/.xsession-errors:
Cjs-message: 09:26:35.004: JS LOG: Ignored exception from dbus method: 
GIO.IOErrorEnum: GDBus:Error:org.gnome.SessionManager.NotInRunning: 
Shutdown interface is only available during the Running phase




OpenPGP_signature
Description: OpenPGP digital signature