Re: What's best way to handle HTML emails in Mutt?

[Andy Smith]

Hi,

On Fri, Dec 27, 2024 at 11:41:39PM -0500, hobie of RMN wrote:
> For 20 years I've enjoyed Mutt as my primary mail reader.  I used its
> 'bounce' feature to deal with HTML mails, sending them to a webmail
> program on a different server.

I also use Mutt. I have Mutt's mailcap set to view the text/html part
with w3m. This is good enough for most cases but where the mail is
complicated I just use a keybinding to load it into a local firefox.

If you are running Mutt on the same computer that your web browser is on
then you could just set up your mailcap to always view with the browser.

If like me you are running Mutt on some other machine it gets a bit more
complicated as you can't just have mutt execute the browser. Even if X
forwarding worked it would be painful. So I call a script from Mutt's
mailcap for text/html that:

1. works out which host I am connected from currently

2. copies the html part to that host

3. connects to that host and executes a program that opens a new tab in
  my running browser, with the path to the html file from (2) as the
  URL.

> now the jellyfish anti-spam daemon won't allow me to do that; it
> rejects my bounced emails because "From:" and "(envelope-from)" are
> not the same.

I'm amazed that this worked for you for as long as it did to be honest,
due to things like SPF and DKIM.

SPF is basically a DNS record that enumerates the IP networks that can
send email "from" a domain. When you bounce an email you're asserting
that you are authorised to send emails "from" users at that domain, but
your domain's SPF record only allows certain hosts.

If that record is under your control, and if your machine with Mutt has
a static IP address, then you could add it to the record.

Otherwise if you can find a way to send yourself the email while
changing the From: address to something else then it will probably
continue to work for you. However, many Email Service Providers do now
require that all incoming email have at least one of SPF or DKIM
passing, so that's an added wrinkle.

Thanks,
Andy

-- 
https://bitfolk.com/ -- No-nonsense VPS hosting

Please consider the environment before reading this e-mail.
 — John Levine

Re: synaptic workalike that WILL run on sudo with wayland.

[Andy Smith]

Hi,

On Fri, Dec 27, 2024 at 08:26:26PM -0500, gene heskett wrote:
> When is that going to happen? Trying to maintain debian-arm stuff blind is a
> pita. I want to SEE whats available.

I have never in my life felt the need to use synaptic. What am I
missing out on?

I just use "apt" from a terminal. apticron emails me which updates are
available.

Running headless servers but still wanting a GUI app on each and every
one of them to manage updates? Doesn't sound like the Linux I know!

Thanks,
Andy

-- 
https://bitfolk.com/ -- No-nonsense VPS hosting

Re: synaptic workalike that WILL run on sudo with wayland.

[Andy Smith]

Hi,

On Fri, Dec 27, 2024 at 10:35:44PM -0600, David Wright wrote:
> I know you won't use aptitude, so why don't you try running
> synaptic on wayland, decide what changes you want, and then
> use sudo apt to actually install/remove the packages.

I think Gene's situation is probably also complicated by the fact that
the host he's running these commands on is not his actual desktop.

It's not clear to me if Gene has a full desktop environment installed on
these Banana Pis with a display/keyboard/mouse attached and is doing
commands locally, or if Gene's running Mate desktop on his "real"
desktop computer and connected to a Banana Pi over SSH from a terminal.

If the latter then do we know on WHICH of those two computers Gene has
installed all this polkit stuff? And does polkit even work over the
network on Wayland like it would on X?

Then of course is the fact that Gene won't be running Debian on the
Banana Pi…

Another challenging thread.

Thanks,
Andy

-- 
https://bitfolk.com/ -- No-nonsense VPS hosting

Re: Firefox lockups with kernel 6.12.6

[George at Clug]

On Saturday, 28-12-2024 at 12:34 John Hasler wrote:
> As an interim solution consider installing the "New Tab Suspender v2"
> extension:

John,

I installed the "New Tab Suspender v2" Add-on but how do you 1) configure the 
timer and the Tab count, 2) know if it is actually working or not?

George


> 
>   A very lightweight tab suspender to suspend inactive tabs that reduces
>   an overall memory usage of firefox, uses a firefox native discard api
> -- 
> John Hasler 
> j...@sugarbit.com
> Elmwood, WI USA
> 
> 

Trixie and Nvidia proprietary drivers

[George at Clug]

​Hi,


After installing updates this morning, my Debian Trixie, KDE X11
installation hangs while loading Trixie with Nvidia proprietary
drivers as packaged with Debian Trixie (the lockup is before the logon
screen appears).


My Nvidia GPU is a RTX 2070. Intel i5 processor.


I did a reboot and logged in maintenance mode, removed the Nvidia
drivers (" # apt purge nvidia* ") which allowed me to log in with
Nouveau drivers. 



When I reinstalled the proprietary Nvidia drivers I had the same issue
of the boot process hanging, and had to go back to Nouveau drivers. 
Which I am currently using.


I am wondering if anyone else has experience this issue?


I expect if I wait long enough someone will have resolved the issue.


I am so curious if and when Debian Trixie will be packaged with the
Nvidia stable version of the NVIDIA 565.77 drivers?  I am looking
forward to this, also hoping they are introduced in Debian Bookworm,
now that the drivers are 'stable'. My guess is that KDE Plasma 6 will
work well with Wayland using these drivers.



https://9to5linux.com/nvidia-565-linux-graphics-driver-is-now-available-for-download-heres-whats-new
NVIDIA released today NVIDIA 565.77 as the first stable version of the
NVIDIA 565 graphics driver series for Linux, FreeBSD, and Solaris
systems with various improvements and multiple bug fixes.


George.

Re: What's best way to handle HTML emails in Mutt?

[Chris Green]

hobie of RMN  wrote:
> For 20 years I've enjoyed Mutt as my primary mail reader.  I used its
> 'bounce' feature to deal with HTML mails, sending them to a webmail
> program on a different server.  But now the jellyfish anti-spam daemon
> won't allow me to do that; it rejects my bounced emails because "From:"
> and "(envelope-from)" are not the same.
> 
> What's the best way to handle this?  Switch to Thunderbird or claws-mail?
> 
I simply use lynx to view 99% of HTML E-Mail and the odd one that
doesn't view well by that means I feed into my web browser.  All on
the same machine.

In ~/.mailcap I have:-

text/html; lynx -dont_wrap_pre -dump %s; copiousoutput; nametemplate=%s.html

That does the work of viewing most HTML E-Mail as text.  Then, when
that doesn't work I use 'v' in mutt to view the parts of the E-Mail
and feed the text/html bit to my browser with 'm' (view-mailcap).

To do this second bit I have the following in ~/.mailcap:-

text/html; /home/chris/bin/muttview %s html

This comes before the line that calls lynx, I can't quite remember how
it works, I think it's the 'copiousoutput' bit that means that line
gets selected by default.  

The /home/chris/bin/muttview is rather complicated in my case because
I run mutt remotely by ssh and feed the html back through the ssh
connection with a reverse tunnel to view in the browser on the client
machine.  If you are running mutt locally then just call your web
browser.

-- 
Chris Green
·

Re: What's best way to handle HTML emails in Mutt?

[Olafur Jens Sigurdsson]

On Sat, Dec 28, 2024 at 10:23:09AM +, Chris Green wrote:
> hobie of RMN  wrote:
> > For 20 years I've enjoyed Mutt as my primary mail reader.  I used its
> > 'bounce' feature to deal with HTML mails, sending them to a webmail
> > program on a different server.  But now the jellyfish anti-spam daemon
> > won't allow me to do that; it rejects my bounced emails because "From:"
> > and "(envelope-from)" are not the same.
> > 
> > What's the best way to handle this?  Switch to Thunderbird or claws-mail?
> > 
> I simply use lynx to view 99% of HTML E-Mail and the odd one that
> doesn't view well by that means I feed into my web browser.  All on
> the same machine.
> 

And if you don't like lynx then w3m might be a better choice.

Cheers, Oli

Re: Firefox lockups with kernel 6.12.6

[John Hasler]

George writes:
> I installed the "New Tab Suspender v2" Add-on but how do you 1) configure
> the timer and the Tab count, 2) know if it is actually working or not?

It doesn't seem to be configureable.  You can tell it's working when
long-unused tabs reload when you click them.  There are several other
tab unloaders such as Auto Tab Discard.

You can also go to about:config, search for "memory", and adjust things
like "browser.cache.memory.capacity" and
"browser.tabs.unloadOnLowMemory".
-- 
John Hasler 
j...@sugarbit.com
Elmwood, WI USA

How to update fonts

[Alain D D Williams]

I am running Debian 12 - Bookworm.

I occasionally see Unicode characters that do not do not display properly. Eg:
메리 크리스마스 (for the curious: this says Happy Christmas in Korean).

These do however display properly on my laptop which runs Mint 21.3.

I suspect that I could see them if I used the testing version of some package.
Which one(s) and how do I do this ?

They fail to display in either Mate terminal 1.26 or in Firefox 128.5 so I
suspect it is a system thing.

Thanks in advance.

-- 
Alain Williams
Linux/GNU Consultant - Mail systems, Web sites, Networking, Programmer, IT 
Lecturer.
+44 (0) 787 668 0256  https://www.phcomp.co.uk/
Parliament Hill Computers. Registration Information: 
https://www.phcomp.co.uk/Contact.html
#include 

Re: that's disturbing

[mick.crane]

On 2024-12-28 17:12, mick.crane wrote:
a couple of months ago I updated Debian from Buster to Bookworm and my 
roundcube/dovecot installation stopped working.
Apache just presented a directory listing and my index.html link to the 
roundcube install displayed a config file.

I thought I should get around to fixing it but put it off.
Just now I thought I'd get a raspberry Pi and reinstall everything on 
that as they are smaller things.

I'll just look one more time
and... it's working.
I didn't do anything, haven't rebooted it.
What could that be?
Browser cache, DNS sorting itself out, some automatic update?
It is disturbing it just started working on its own.
mick


Direct upgrades from N to N+2 is not supported. You ought to have first 
upgraded to Bullseye then to Bookworm. If you have a backup of the 
Buster image you could go that route and it should work, unless you have 
some "non-standard" things installed which would be documented in both 
the Bullseye and Bookworm release notes.


I tried this in the past and it blew up spectacularly...then I read the 
documentation.

-
Sorry that's me not making the effort to check which of the "B"s it was. 
Would have been the previous one, Bullseye

Now I should sort out using the right SMTP severs.
mick.

Re: How to update fonts

[Nicolas George]

Alain D D Williams (12024-12-28):
> I suspect that I could see them if I used the testing version of some package.
> Which one(s) and how do I do this ?

I strongly doubt a common font will be “upgraded” with Korean
characters. More likely, the package is just not installed.

So: find out which font and which file contains it, query the package it
belongs to. At worse, you can use strace to see what files the program
opens; possibly easier, in /proc/$PID/map_files/ you can see which font
files are currently in use.

Regards,

-- 
  Nicolas George

Re: synaptic workalike that WILL run on sudo with wayland.

[David Wright]

On Sat 28 Dec 2024 at 01:07:58 (-0500), gene heskett wrote:
> On 12/27/24 23:36, David Wright wrote:
> > On Fri 27 Dec 2024 at 20:26:26 (-0500), gene heskett wrote:
> > > When is that going to happen? Trying to maintain debian-arm stuff
> > > blind is a pita. I want to SEE whats available.
> > I know you won't use aptitude, so why don't you try running
> > synaptic on wayland, decide what changes you want, and then
> > use sudo apt to actually install/remove the packages.
> I would do that, except highlight/copypaste does not work with either
> agent.  And at 90 yo, I have a hellofatime remembering 40+ character
> command lines.

Perhaps you're unaware that bash-completion will complete package
names for you, in a similar way to filename completion. So long as
you don't think that mate-polkit is called policykit-mate, you should
be able to reduce the number of characters needing to be typed.

As for remembering, I was under the impression that wayland could
display more than one window on the screen at the same time. I find
it difficult to believe that that's not true.

Do remember, however, that you'll need to refresh the synaptic screen
whenever you install/remove packages, because it won't otherwise take
account of the new system state.

Cheers,
David.

Re: What's best way to handle HTML emails in Mutt?

[Henning Follmann]

Just a few additional bits added

On Sat, Dec 28, 2024 at 10:23:09AM +, Chris Green wrote:
> hobie of RMN  wrote:
> > For 20 years I've enjoyed Mutt as my primary mail reader.  I used its
> > 'bounce' feature to deal with HTML mails, sending them to a webmail
> > program on a different server.  But now the jellyfish anti-spam daemon
> > won't allow me to do that; it rejects my bounced emails because "From:"
> > and "(envelope-from)" are not the same.
> > 
> > What's the best way to handle this?  Switch to Thunderbird or claws-mail?
> > 
> I simply use lynx to view 99% of HTML E-Mail and the odd one that
> doesn't view well by that means I feed into my web browser.  All on
> the same machine.
>
To get most HTML displayed inside mutt you could add this to your .muttrc:

# for html mails
alternative_order text/plain text/html
unauto_view *
auto_view text/html
macro index,pager \cb " urlscan" "call urlscan to extract 
URLs out of a message"
macro attach,compose \cb " urlscan" "call urlscan to extract 
URLs out of a message"

The bottom two lines are for urlscan. The output is not really pretty, but
you will see the real url . These then will open with whatever xdg-open is
set to. 

> In ~/.mailcap I have:-
> 
> text/html; lynx -dont_wrap_pre -dump %s; copiousoutput; 
> nametemplate=%s.html
> 

As some pointed out the other option w3m;
This is what is in my .mailcap:

text/html; w3m -dump -o display_link_number=true -I %{charset} -T text/html 
'%s'; copiousoutput; description=HTML Text; %nametemplate=%s.html; 
needsterminal; priority=1


> That does the work of viewing most HTML E-Mail as text.  Then, when
> that doesn't work I use 'v' in mutt to view the parts of the E-Mail
> and feed the text/html bit to my browser with 'm' (view-mailcap).
[...]



-H

-- 
Henning Follmann   | hfollm...@itcfollmann.com

Re: synaptic workalike that WILL run on sudo with wayland.

[Greg Wooledge]

On Sat, Dec 28, 2024 at 12:53:41 +, Darac Marjal wrote:
> You can run X clients on the server and remotely display that on a
> workstation (i.e. something _with_ a head), but AFAIK the wayland protocol
> doesn't support any sort of network transport (you can run a GUI application
> on a Wayland display by using Xwayland, but that's sort of by-the-by).

Is this really true?  And yet, they expect everyone to switch from X to
Wayland, when this is true?

No thank you.

Correct way to stop fetchmail

[Roger Price]

I would like to stop fetchmail on Debian 11.  I tried command systemctl stop 
fetchmail and system status fetchmail reported


 ● fetchmail.service - LSB: init-Script for system wide fetchmail daemon
 Loaded: loaded (/etc/init.d/fetchmail; generated)
 Active: inactive (dead) since Sat 2024-12-28 10:37:58 CET; 6h ago
   Docs: man:systemd-sysv-generator(8)
Process: 1250 ExecStart=/etc/init.d/fetchmail start (code=exited, 
status=0/SUCCESS)
Process: 1199929 ExecStop=/etc/init.d/fetchmail stop (code=exited, 
status=0/SUCCESS)
  Tasks: 1 (limit: 57750)
 Memory: 1.9M
CPU: 6min 33.392s
 CGroup: /system.slice/fetchmail.service
 └─1264 /usr/bin/fetchmail -f /etc/fetchmarc
--pidfile /var/run/fetchmail/fetchmail.pid --syslog
 ...
 Dec 28 17:06:54 titan fetchmail[1264]: reading message
mail...@rogerprice.org@mail.gandi.net:1 
of 1 (9321 octets) flushed

It's dead, but still fetching mail! I tried

 root@titan ~ /etc/init.d/fetchmail --quit
 Not starting fetchmail daemon, disabled via /etc/default/fetchmail.

but fetchmail goes on fetching mail.  What is the correct way of stopping 
fetchmail?


Roger

Re: synaptic workalike that WILL run on sudo with wayland.

[gene heskett]

On 12/28/24 09:43, Greg Wooledge wrote:

On Sat, Dec 28, 2024 at 12:53:41 +, Darac Marjal wrote:

You can run X clients on the server and remotely display that on a
workstation (i.e. something _with_ a head), but AFAIK the wayland protocol
doesn't support any sort of network transport (you can run a GUI application
on a Wayland display by using Xwayland, but that's sort of by-the-by).

Is this really true?  And yet, they expect everyone to switch from X to
Wayland, when this is true?

No thank you.

.

+1000

Cheers, Gene Heskett, CET.
--
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis

Re: synaptic workalike that WILL run on sudo with wayland.

[gene heskett]

On 12/28/24 08:24, Andy Smith wrote:

Hi,

On Sat, Dec 28, 2024 at 12:53:41PM +, Darac Marjal wrote:

On 28/12/2024 10:06, Andy Smith wrote:

On Fri, Dec 27, 2024 at 08:26:26PM -0500, gene heskett wrote:

When is that going to happen? Trying to maintain debian-arm stuff blind is a
pita. I want to SEE whats available.

I have never in my life felt the need to use synaptic. What am I
missing out on?

Synaptic is a GUI frontend to apt. It's not really much different to
aptitude, except that it's maybe a bit more mouse-friendly.

That much I know. I am wondering what its killer feature is that Gene
wants to work so hard to make it run on a headless single board computer
that one would think to be wholly unsuitable for running GUI apps.


armbian in the full desktop version runs noticeably slower than it runs 
on amd64 stuff, but its more than fast enough to run a full screen gui 
for a 3d printer.  I'll gladly tolerate a just noticeable lag that never 
crashes, uptimes from kernel update to kernel update, with 
klipper/moonraker/fluidd feeding nginx, broadcasting to any web browser 
in the house that is watching aliasname:80, in exchange for an 18 watt 
power draw (including its 24" AOL monitor) while sitting idle between 
jobs. I use firefox to localhost:80 on the local machine actually 
running the printer at speeds around 10x what it could do OOTB 4 years 
ago when it was new.


FWIW, I probably have another $2000 in better hdwe., much of which I 
designed in OpenSCAD. I may ask what seems to be off the wall questions 
simply hoping to glean additional info because I enjoy plowing new 
ground and making it work better.  Part of the upgrade is closed loop 
stepper/servos, because they have at least a 10x improvement in the 
accuracy of motor control. The error at the instant determines motor 
current, running at 7x speeds, average motor power is 30% of the burn 
your hand temps the usual setup does, it can even reverse a motor that 
overshoots, putting it back where it belongs using around 4x the instant 
power if it has to. The latest addition is a motorized nozzle cleaner, 
operated as part of the startup gcode.


I've got to get my web page up again, I lost it when those two seagate 
2T drives went offline at 2 weeks accumulated runtime as I was trying 
make bookworm run now 18 months ago. It runs when it gets around to it, 
opening a local file is an automatic lockup for 30 seconds. And no one 
has identified that problem. Around 30 fresh installs with the only usb 
plugged in being the wireless button for logitek keyboard/mouse but I 
still get orca and brltty yelling every keystroke at me thru my 
speakers. Yet when I question the broken bookworm installer, I catch 
hell from the powers that be here. Total denial that the problem exists 
has been the universal response.





I just use "apt" from a terminal. apticron emails me which updates are
available.

Running headless servers but still wanting a GUI app on each and every
one of them to manage updates? Doesn't sound like the Linux I know!

You can't run a GUI on a headless server. That's the whole point of it
being headless. You can run X clients on the server and remotely
display that on a workstation (i.e. something _with_ a head)

I thought it was obvious that was what I was referring to, but clearly
not.

Even that makes no sense to me in most cases, certainly not in the case
of multiple Banana Pi and a desktop that seems to run Wayland. Hence I
am suggesting that Gene is going in completely the wrong direction here.

Thanks,
Andy


Cheers, Gene Heskett, CET.
--
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis

Re: Correct way to stop fetchmail

[Roger Price]

Thanks for your replies.  I tried

 systemctl stop fetchmail.service

I tried 4 times. then I tried

 /etc/init.d/fetchmail stop

but although dead for 6 hours, fetchmail continued to fetch mails.
So I tried

 root@titan ~ ps -ef | grep fetch
 fetchma+1264   1  0 Sep25 ?00:06:33 /usr/bin/fetchmail -f
  /etc/fetchmailrc --pidfile /var/run/fetchmail/fetchmail.pid --syslog
 root@titan ~ kill 1264

And the fetchma+ process disappeared - I hope it no longer rises from the 
dead.


Roger

Re: that's disturbing

[John Boxall]

Direct upgrades from N to N+2 is not supported. You ought to have first
upgraded to Bullseye then to Bookworm. If you have a backup of the Buster
image you could go that route and it should work, unless you have some
"non-standard" things installed which would be documented in both the
Bullseye and Bookworm release notes.

I tried this in the past and it blew up spectacularly...then I read the
documentation.

On Sat, Dec 28, 2024 at 1:07 PM mick.crane  wrote:

> a couple of months ago I updated Debian from Buster to Bookworm and my
> roundcube/dovecot installation stopped working.
> Apache just presented a directory listing and my index.html link to the
> roundcube install displayed a config file.
> I thought I should get around to fixing it but put it off.
> Just now I thought I'd get a raspberry Pi and reinstall everything on
> that as they are smaller things.
> I'll just look one more time
> and... it's working.
> I didn't do anything, haven't rebooted it.
> What could that be?
> Browser cache, DNS sorting itself out, some automatic update?
> It is disturbing it just started working on its own.
> mick
>
>

-- 
Regards,

John Boxall

Re: What's best way to handle HTML emails in Mutt?

[Greg Wooledge]

On Fri, Dec 27, 2024 at 23:41:39 -0500, hobie of RMN wrote:
> For 20 years I've enjoyed Mutt as my primary mail reader.  I used its
> 'bounce' feature to deal with HTML mails, sending them to a webmail
> program on a different server.  But now the jellyfish anti-spam daemon
> won't allow me to do that; it rejects my bounced emails because "From:"
> and "(envelope-from)" are not the same.
> 
> What's the best way to handle this?

It depends on what you want to do.  For me, there are two different
modes of dealing with HTML email in mutt.

The first mode is allowing mutt to run a text web browser to render
the HTML into readable text.  It's been a while since I configured
this, but I believe it's these three lines in my .muttrc which do it:

alternative_order text/plain text/html
unauto_view *
auto_view = text/html

If a message is sent in mixed mode (text/plain AND text/html parts),
then mutt shows the text/plain part.  (Note: this may or may not be
a faithful rendering of the HTML part.  It's whatever the sender put
in that part, and it may be missing some information.)

If a message is sent with ONLY a text/html part, then mutt chooses
a text web browser to render it.  I recommend installing "links"
because it does a better job than "lynx".  If links is not present
but lynx is, lynx will be used, which is... marginally acceptable.

The second mode that I use is to copy the HTML part to a file, and
then scp it to my local machine, and then load it in a web browser.
While reading the message, press 'v' to view the parts in a tree
layout.  Use the arrows to select the text/html part, then press 's'
to save it to a file.  You'll be prompted for a filename; I use the
same filename every time (foo.html) just to make the procedure easy.

Once foo.html is saved on the mutt host, scp it to your local host.
If your local and remote hosts are the same machine, then you can
skip this step.  Of course, you can use rsync, sftp, or whatever.

Once foo.html is copied to the local host, go to your web browser
(Firefox or Chromium or whatever), press Ctrl-t to open a new tab,
and select file:///home/yourname/foo.html in the URL bar.  If you
use the same filename every time, then this will be in your history
already (after the first time), which greatly reduces the amount of
typing needed in this step.  Hit Enter, and voila.  Now you're
reading the text/html part the way the sender intended, including
loading all of the tracking images that tell them you've opened the
email.

Now, this is just my personal workflow, and you may come up with
different ways to deal with it.  Just find whatever works best for
you.

that's disturbing

[mick.crane]

a couple of months ago I updated Debian from Buster to Bookworm and my 
roundcube/dovecot installation stopped working.
Apache just presented a directory listing and my index.html link to the 
roundcube install displayed a config file.

I thought I should get around to fixing it but put it off.
Just now I thought I'd get a raspberry Pi and reinstall everything on 
that as they are smaller things.

I'll just look one more time
and... it's working.
I didn't do anything, haven't rebooted it.
What could that be?
Browser cache, DNS sorting itself out, some automatic update?
It is disturbing it just started working on its own.
mick

Re: synaptic workalike that WILL run on sudo with wayland.

[Lee]

On Sat, Dec 28, 2024 at 1:08 AM gene heskett wrote:
>
>
> On 12/27/24 23:36, David Wright wrote:
> > On Fri 27 Dec 2024 at 20:26:26 (-0500), gene heskett wrote:

> >> When is that going to happen? Trying to maintain debian-arm stuff
> >> blind is a pita. I want to SEE whats available.

> > I know you won't use aptitude, so why don't you try running
> > synaptic on wayland, decide what changes you want, and then
> > use sudo apt to actually install/remove the packages.

> I would do that, except highlight/copypaste does not work with either
> agent.

Oh?  Have you tried running synaptic, getting the list of packages to
install / update, right clicking on the package name and selecting
properties?
That has a package name that I can select with the mouse and copy the
name to the clipboard
Then on your terminal run 'sudo apt install ' and have apt install the package.

>  And at 90 yo, I have a hellofatime remembering 40+ character
> command lines.

Copy/paste is a requirement for me also.  If your terminal program
doesn't do what you want, take a look at the current version of
xfce4-terminal - that has most of what I was looking for.

Regards,
Lee

Re: Correct way to stop fetchmail

[Charles Curley]

On Sat, 28 Dec 2024 17:21:08 +0100 (CET)
Roger Price  wrote:

> but fetchmail goes on fetching mail.  What is the correct way of
> stopping fetchmail?

Perhaps there is a fetchmail.timer unit that is starting up
fetchmail.service from time to time?

systemctl status fetchmail.timer

You might also look at /usr/share/doc/fetchmail/README.systemd

-- 
Does anybody read signatures any more?

https://charlescurley.com
https://charlescurley.com/blog/

Re: Correct way to stop fetchmail

[Greg Wooledge]

On Sat, Dec 28, 2024 at 11:36:44 -0700, Charles Curley wrote:
> On Sat, 28 Dec 2024 17:21:08 +0100 (CET)
> Roger Price  wrote:
> 
> > but fetchmail goes on fetching mail.  What is the correct way of
> > stopping fetchmail?
> 
> Perhaps there is a fetchmail.timer unit that is starting up
> fetchmail.service from time to time?
> 
> systemctl status fetchmail.timer
> 
> You might also look at /usr/share/doc/fetchmail/README.systemd

Going back to the original post, it looks like the systemd "unit"
for fetchmail is actually just a shim for the init.d scripts.
That may change in some future version of Debian (or perhaps the OP
is on an old version??), but for now, I'd look to more traditional
sources.

Perhaps there's a cron job that starts it.

Assuming "fetchmail" is the package name, it might be worth looking at
the file list (dpkg -L fetchmail) to see what's being installed.  If
there's a *.timer file, or a file inside a cron.* directory, those
would be worth investigating.

Re: Correct way to stop fetchmail

[Greg Wooledge]

On Sat, Dec 28, 2024 at 18:56:09 +0100, Roger Price wrote:
> I tried 4 times. then I tried
> 
>  /etc/init.d/fetchmail stop
> 
> but although dead for 6 hours, fetchmail continued to fetch mails.
> So I tried
> 
>  root@titan ~ ps -ef | grep fetch
>  fetchma+1264   1  0 Sep25 ?00:06:33 /usr/bin/fetchmail -f
>   /etc/fetchmailrc --pidfile /var/run/fetchmail/fetchmail.pid --syslog
>  root@titan ~ kill 1264

That process had been running for three months.(*)  It wasn't "dead for 6
hours" at the moment you ran those commands.

(*) Unless of course your system clock is massively wrong.

Re: Correct way to stop fetchmail

[Greg Wooledge]

On Sat, Dec 28, 2024 at 17:21:08 +0100, Roger Price wrote:
>  ● fetchmail.service - LSB: init-Script for system wide fetchmail daemon

The correct way to stop a systemd service whose name you know is:

systemctl stop fetchmail.service

If you're entirely new to systemd, you can get by with:

systemctl status NAME   # looks like you found that
systemctl stop NAME # terminate a running service
systemctl start NAME# start a stopped service
systemctl restart NAME  # stop, and then start, a running service
systemctl reload NAME   # ask a running service to reread configs

Not all services have a "reload" option, but many do.

Quite often, discovering the service's systemd unit name is the hardest
part in managing it.

> It's dead, but still fetching mail!

If the service is showing as terminated, but it continues running
intermittently, then it's *possible* that something else is invoking
it.  In that case, you will have some detective work to do.  But let's
not go down that rabbit hole if it's just a matter of you not knowing
how to stop a service.

> I tried
> 
>  root@titan ~ /etc/init.d/fetchmail --quit
>  Not starting fetchmail daemon, disabled via /etc/default/fetchmail.

The correct way to use a legacy sysv-rc init.d script to stop a service
is:

/etc/init.d/fetchmail stop

But if you already know there's a systemd unit, you should use the
systemctl way instead.  The init.d scripts are remnants of a past
init system.  If you're not actually *using* that init system, you
shouldn't get in the habit of relying on its legacy shims to be there.

In this specific case, the --quit option you tried to use was ignored,
and it appears to have assumed you meant "start".  But it DID NOT
start the fetchmail service, because it's disabled in the
/etc/default/fetchmail file.  (The /etc/default/ directory is the
common place for configuration files used by sysv-rc scripts.)

All of this is just background noise, though.  You shouldn't be using
this init.d script when there's a systemd unit file.

Re: synaptic workalike that WILL run on sudo with wayland.

[Greg Wooledge]

On Sat, Dec 28, 2024 at 12:13:18 -0500, Lee wrote:
> Copy/paste is a requirement for me also.  If your terminal program
> doesn't do what you want, take a look at the current version of
> xfce4-terminal - that has most of what I was looking for.

One thing that *might* be relevant here is that X11 uses three separate
paste buffers, called "primary", "secondary" and "clipboard".  When
you attempt to paste something, depending on what application you're
pasting into and what *method* you're using to paste, you might be
pulling from a different paste buffer than you want.

Also, the application you're pasting *from* may have put the content
that you want to paste into a different buffer than you expected.

Your terminal (or other application being pasted into) may support
multiple methods of buffer selection.  For example,

* Right mouse button: usually pastes from "primary".
* Shift+Insert keys: usually pastes from "clipboard".

You can generally ignore "secondary".  I don't know of anything that
uses it.

If you install the xclip package, you can use that to see what's in
each paste buffer.

xclip -o -selection {primary|secondary|clipboard} | less

You can also use it to "transfer" content to another paste buffer, if
you need that for your terminal/application.

xclip -o -selection clipboard | xclip -i# primary is default

I don't know anything about synaptic, and have no desire to learn it,
but if you're having trouble pasting content out of it, then perhaps
it's simply using the "wrong" paste buffer.

I also don't know anything about Wayland, so everything here might not
apply in Wayland.

Re: How to update fonts

[Alain D D Williams]

On Sun, Dec 29, 2024 at 09:09:46AM +0700, Max Nikulin wrote:

> Do you have the "fonts-recommended" package installed?

That did the trick thanks - then a reboot. I might have been able to get away
with a logout.

I also now seem to be able to see all the emojis that my daughter sends me.

Thanks

> You may skim through
> 
> apt search font korean
> apt list '~nfont~i'
> 

-- 
Alain Williams
Linux/GNU Consultant - Mail systems, Web sites, Networking, Programmer, IT 
Lecturer.
+44 (0) 787 668 0256  https://www.phcomp.co.uk/
Parliament Hill Computers. Registration Information: 
https://www.phcomp.co.uk/Contact.html
#include 

Re: How to update fonts

[Max Nikulin]

On 29/12/2024 02:28, Alain D D Williams wrote:


I occasionally see Unicode characters that do not do not display properly. Eg:
메리 크리스마스 (for the curious: this says Happy Christmas in Korean).

[...]

They fail to display in either Mate terminal 1.26 or in Firefox 128.5 so I
suspect it is a system thing.


I have no idea if Thunderbird renders Korean glyph properly, but failure 
(if any) at least is not apparent to me. I expect that Gecko is able to 
find a fallback font if it is installed.


Do you have the "fonts-recommended" package installed?

You may skim through

apt search font korean
apt list '~nfont~i'

Re: double-boot laptop with Debian Trixie

[Charles Curley]

On Sun, 29 Dec 2024 01:20:17 +0100
Rafał Lichwała  wrote:

> Hi,
> 
> I plan to switch completely from Windows 11 to Debian on my laptop,
> but as a first step I'd like to configure double-boot setup to give
> it a try, look around and check if all hardware works fine under
> Debian.
> 
> Current state:
> Dell laptop (one NVMe disk only) with Windows 11 onboard.
> Windows starts in UEFI mode (Secure Boot ON).
> Already resized main partition to make space for Debian.
> So currently there are the following partitions on disk:
> 
> 100MB, FAT32, GPT (EFI System Partition)
> 16MB, Other, GPT (Reserved Partition)
> 300GB, NTFS, GPT (Data Partition) (Main C: disk visible under Win11)
> 168GB, free space, ready to have ~150GB for Debian and 16GB for swap
> 880MB, NTFS, GPT (Recovery Partition)
> 
> First try:
> UEFI Boot (Secure Boot ON), Debian Trixie,

Is there any reason for trixie? Both trixie and its installer (d-i) are
evolving and may not give you useful results. I suggest you use the
current Stable, Debian 12.8, bookworm.

> Expert mode installation During installation Debian does NOT see my
> disk at all (pendrive with Debian Trixie is visible only)

Interesting. I recently installed Debian 12 to an NVMe machine with no
issues.

> Also during installation Debian says that "it seems there is another 
> system installed in Legacy Mode, am I  sure to force UEFI mode?"
> which is not correct - Win11 is installed in UEFI mode (when I try to
> boot it in Legacy Mode I get Windows Blue Screen Of Death)

Have you changed this in the firmware (formerly known as the BIOS)?
Make sure it is set for UEFI booting. Secure boot should probably be
off, but you should be able to turn it back on when you are done.


> 
> Second try:
> Secure Boot set to OFF.
> Nothing changed during Trixie installation, so no device to install
> on.
> 
> Third try:
> Boot in Legacy Mode (so Secure Boot is OFF).
> Debian Trixie can see my laptop disk properly, so installation is 
> successful, but...
> It cannot see my Windows 11 installation when trying to 
> install/configure/prepare GRUB.
> At this step I decided to NOT install GRUB on my main disk to NOT
> loose access to my Windows 11 system.
> (but Trixie says that it's possible to restore access by manual 
> configuration of GRUB later on - I'm not sure...)
> 
> Main goal:
> To have both current Windows 11 and new Debian Trixie available in
> GRUB and properly bootable.

I take it "properly bootable" means UEFI, and secure boot optional.

> 
> How to achieve this? :-)

I'd start by trying to install Debian 12, bookworm.


-- 
Does anybody read signatures any more?

https://charlescurley.com
https://charlescurley.com/blog/

Re: Correct way to stop fetchmail

[Max Nikulin]

On 28/12/2024 23:47, Greg Wooledge wrote:

On Sat, Dec 28, 2024 at 17:21:08 +0100, Roger Price wrote:

  root@titan ~ /etc/init.d/fetchmail --quit
  Not starting fetchmail daemon, disabled via /etc/default/fetchmail.


The correct way to use a legacy sysv-rc init.d script to stop a service
is:

 /etc/init.d/fetchmail stop


I do not have fetchmail installed, so I am unaware of its idiosyncrasy. 
At certain step, most of init.d scripts were modified to use systemctl 
commands when systemd is running, so usually it should not matter if a 
systemd unit generated from its SysV init script is started or stopped 
using "service", "systemctl", or direct script invocation.


When properly configured, systemd unit should properly stop if the 
daemon exits with 0 exit code e.g. in response to some direct command.


My guess is that the main service process exited, but some its child (it 
belongs to service's cgroup) were continuing running for some reason. 
Unless improperly configured, systemd is able to kill all processes 
after some timeout. Perhaps something is wrong with service type, 
KillMode, or timeouts.


Commands that might help to discover what actually happened

systemctl cat fetchmail.service
journalctl -u fetchmail.service

Re: Correct way to stop fetchmail

[Greg Wooledge]

On Sun, Dec 29, 2024 at 09:40:44 +0700, Max Nikulin wrote:
> On 28/12/2024 23:47, Greg Wooledge wrote:
> > On Sat, Dec 28, 2024 at 17:21:08 +0100, Roger Price wrote:
> > >   root@titan ~ /etc/init.d/fetchmail --quit
> > >   Not starting fetchmail daemon, disabled via /etc/default/fetchmail.
> > 
> > The correct way to use a legacy sysv-rc init.d script to stop a service
> > is:
> > 
> >  /etc/init.d/fetchmail stop
> 
> I do not have fetchmail installed, [...]

Yeah, that's the common problem here.  Only the OP is using this package,
and apparently he's not on the current stable release.

> My guess is that the main service process exited, but some its child (it
> belongs to service's cgroup) were continuing running for some reason. Unless
> improperly configured, systemd is able to kill all processes after some
> timeout. Perhaps something is wrong with service type, KillMode, or
> timeouts.

Right.  I had a similar guess (parent dies, children live on).

I looked at the file list for the oldstable package (the one I believe
Roger is using) at
 and
this one does NOT have a systemd unit file.  As I've pointed out a
couple times now, that version of the package only has an init.d file.

So, there is no service type, there is no KillMode, and I'm not *aware*
of any kind of timeouts enforced by the init.d script.  But I haven't
actually read it.

Anyway, I'm still going with the "init.d script failure, made irrelevant
in the stable release because now it has a systemd unit file" theory.
There's not much point trying to debug an init.d script in an oldstable
release, when the stable release won't be using it (normally).

Re: Correct way to stop fetchmail

[John Hasler]

On Unstable:

toncho/~ 20 systemctl cat fetchmail.service
# /run/systemd/generator.late/fetchmail.service
# Automatically generated by systemd-sysv-generator

[Unit]
Documentation=man:systemd-sysv-generator(8)
SourcePath=/etc/init.d/fetchmail
Description=LSB: init-Script for system wide fetchmail daemon
Before=multi-user.target
Before=multi-user.target
Before=multi-user.target
Before=graphical.target
After=network-online.target
After=remote-fs.target
After=mail-transport-agent.target
After=postfix.service
After=exim4.service
After=nss-lookup.target
Wants=network-online.target

[Service]
Type=forking
Restart=no
TimeoutSec=5min
IgnoreSIGPIPE=no
KillMode=process
GuessMainPID=no
RemainAfterExit=yes
SuccessExitStatus=5 6
ExecStart=/etc/init.d/fetchmail start
ExecStop=/etc/init.d/fetchmail stop

-- 
John Hasler 
j...@sugarbit.com
Elmwood, WI USA

Re: double-boot laptop with Debian Trixie

[George at Clug]

On Sunday, 29-12-2024 at 11:20 Rafał Lichwała wrote:
> Hi,
> 
> I plan to switch completely from Windows 11 to Debian on my laptop, but 
> as a first step I'd like to configure double-boot setup to give it a 
> try, look around and check if all hardware works fine under Debian.
> 
> Current state:
> Dell laptop (one NVMe disk only) with Windows 11 onboard.
> Windows starts in UEFI mode (Secure Boot ON).
> Already resized main partition to make space for Debian.
> So currently there are the following partitions on disk:
> 
> 100MB, FAT32, GPT (EFI System Partition)
> 16MB, Other, GPT (Reserved Partition)
> 300GB, NTFS, GPT (Data Partition) (Main C: disk visible under Win11)
> 168GB, free space, ready to have ~150GB for Debian and 16GB for swap
> 880MB, NTFS, GPT (Recovery Partition)
> 
> First try:
> UEFI Boot (Secure Boot ON), Debian Trixie, Expert mode installation
> During installation Debian does NOT see my disk at all (pendrive with 
> Debian Trixie is visible only)
> Also during installation Debian says that "it seems there is another 
> system installed in Legacy Mode, am I  sure to force UEFI mode?" which 
> is not correct - Win11 is installed in UEFI mode (when I try to boot it 
> in Legacy Mode I get Windows Blue Screen Of Death)
> 
> Second try:
> Secure Boot set to OFF.
> Nothing changed during Trixie installation, so no device to install on.
> 
> Third try:
> Boot in Legacy Mode (so Secure Boot is OFF).
> Debian Trixie can see my laptop disk properly, so installation is 
> successful, but...
> It cannot see my Windows 11 installation when trying to 
> install/configure/prepare GRUB.
> At this step I decided to NOT install GRUB on my main disk to NOT loose 
> access to my Windows 11 system.
> (but Trixie says that it's possible to restore access by manual 
> configuration of GRUB later on - I'm not sure...)
> 
> Main goal:
> To have both current Windows 11 and new Debian Trixie available in GRUB 
> and properly bootable.

Rafal,

I could give better comments if I knew the hardware, particularly the GPU type. 
If I knew the model number I could look up YouTube to see if memory and/or NVMe 
were easily upgradeable, something I confirmed before purchasing my most recent 
laptop.

These are my thoughts, based on my past experiences:

1) Boot Debian live  DVD and check hardware works (my laptop's fingerprint 
reader is not recognised, but that is the only item that does not work). Is it 
possible to use Nvidia proprietary drivers when using a live DVD? I would not 
expect so. I could not get a decent idea of graphics performance for an Nvidia 
GPU without using Nvidia proprietary drivers .

2) If the NVMe was easily replaceable, I would simply put the Windows NVMe 
aside and install a new NVMe for Linux only. This is just my personal 
preference.

3) (as others have suggested) I would recommend installing Debian Bookworm. 
Once all was working well, if you truly had a need for Debian Testing (e.g. 
Debian Trixie), it is not hard to upgrade from Debian Bookworm to Debian 
Trixie.  I expect Debian Trixie will be released in around six months time, but 
until then Debian Trixie is "Testing" and hence could break too easily, like my 
current Debian Trixie installation did yesterday.

4) If you want to dual boot, then I do not see issues using Debian Bookworm to 
do this. I would boot into Debian (i.e. Grub), from where I select either 
Windows or Debian. If my primary use is Windows, then I edit grub to boot to 
Windows first. I have not had issues with UEFI lately, I believe Debian handles 
this well.

5) I use Debian Bookworm as I like stability. Testing means things will break. 
Trixie is currently in testing. I rarely need the latest software. I sometimes 
use backports.
https://backports.debian.org/Instructions/
As a matter of Backports policy, packages in the stable-backports suite are 
taken from Debian testing; however, in rare cases such as security updates, 
packages are taken from unstable. 

6) What GUI have you decided to use?  While I use XFCE, I recommend KDE to 
people, particularly when moving from Windows to Linux. In fact KDE is so nice 
now, maybe I will start using KDE too? (I want both, and I do use both, but on 
different computers)

7) While I like and use Debian, I have recently learned about KDE Linux. 
https://community.kde.org/KDE_Linux

I hope my comments helps you when trying Linux/Debian.

George.


> 
> How to achieve this? :-)
> 
> Thank you,
> Best Regards,
> Rafal
> 
> 
> 

Re: synaptic workalike that WILL run on sudo with wayland.

[tomas]

On Sat, Dec 28, 2024 at 07:46:09PM -0500, gene heskett wrote:
> 
> On 12/28/24 16:41, Andrew M.A. Cater wrote:

[...]

> > Armbian is *NOT* Debian  They do things differently there [...]

> Then you are missing out on the many things a pi clone can do on 5% of an
> amd64's power budget. Granted the amd64 can do it 20x faster, ignoring that
> the pi clone is fast enough. Debian seems to treat the arm64's a toy.

Gene, this is a totally irrelevant answer to what Andrew said. Raspi
might be the best invention since sliced bread and still, asking for
Armbian advice in a Debian group is at least... fraught.

BTW, there is a semi-official Debian for Raspberry Pi, if you care.
Getting a custom kernel up and running on it (as you seem to need for
your lathe work) might be easier or more difficult than on Armbian.

YMMV.

Cheers
-- 
tomás


signature.asc
Description: PGP signature

Re: synaptic workalike that WILL run on sudo with wayland.

[gene heskett]

On 12/29/24 01:30, to...@tuxteam.de wrote:

On Sat, Dec 28, 2024 at 07:46:09PM -0500, gene heskett wrote:

On 12/28/24 16:41, Andrew M.A. Cater wrote:

[...]


Armbian is *NOT* Debian  They do things differently there [...]

Then you are missing out on the many things a pi clone can do on 5% of an
amd64's power budget. Granted the amd64 can do it 20x faster, ignoring that
the pi clone is fast enough. Debian seems to treat the arm64's a toy.

Gene, this is a totally irrelevant answer to what Andrew said. Raspi
might be the best invention since sliced bread and still, asking for
Armbian advice in a Debian group is at least... fraught.

BTW, there is a semi-official Debian for Raspberry Pi, if you care.
Getting a custom kernel up and running on it (as you seem to need for
your lathe work) might be easier or more difficult than on Armbian.


I'm running one, about a year old now, on an actual rpi4b to run my 
biggest lathe, but the last arm64 installer can't find its net interface 
even if you fill in the blanks.  Armbian Just Works. Why can't debian-arm?


Cheers Tomas, Gene Heskett, CET.

--
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis

Re: synaptic workalike that WILL run on sudo with wayland.

[tomas]

On Sun, Dec 29, 2024 at 01:56:06AM -0500, gene heskett wrote:

[...]

> I'm running one, about a year old now, on an actual rpi4b to run my biggest
> lathe, but the last arm64 installer can't find its net interface even if you
> fill in the blanks.  Armbian Just Works. Why can't debian-arm?

I haven't the hardware, so I can't find out.

Cheers
-- 
t


signature.asc
Description: PGP signature

Re: synaptic workalike that WILL run on sudo with wayland.

[Andy Smith]

Hi,

On Sat, Dec 28, 2024 at 12:53:41PM +, Darac Marjal wrote:
> 
> On 28/12/2024 10:06, Andy Smith wrote:
> > On Fri, Dec 27, 2024 at 08:26:26PM -0500, gene heskett wrote:
> > > When is that going to happen? Trying to maintain debian-arm stuff blind 
> > > is a
> > > pita. I want to SEE whats available.
> > I have never in my life felt the need to use synaptic. What am I
> > missing out on?
> Synaptic is a GUI frontend to apt. It's not really much different to
> aptitude, except that it's maybe a bit more mouse-friendly.

That much I know. I am wondering what its killer feature is that Gene
wants to work so hard to make it run on a headless single board computer
that one would think to be wholly unsuitable for running GUI apps.

> > I just use "apt" from a terminal. apticron emails me which updates are
> > available.
> > 
> > Running headless servers but still wanting a GUI app on each and every
> > one of them to manage updates? Doesn't sound like the Linux I know!
> 
> You can't run a GUI on a headless server. That's the whole point of it
> being headless. You can run X clients on the server and remotely
> display that on a workstation (i.e. something _with_ a head)

I thought it was obvious that was what I was referring to, but clearly
not.

Even that makes no sense to me in most cases, certainly not in the case
of multiple Banana Pi and a desktop that seems to run Wayland. Hence I
am suggesting that Gene is going in completely the wrong direction here.

Thanks,
Andy

-- 
https://bitfolk.com/ -- No-nonsense VPS hosting

Re: synaptic workalike that WILL run on sudo with wayland.

[Todd Zullinger]

Greg Wooledge wrote:
> On Sat, Dec 28, 2024 at 12:53:41 +, Darac Marjal wrote:
>> You can run X clients on the server and remotely display that on a
>> workstation (i.e. something _with_ a head), but AFAIK the wayland protocol
>> doesn't support any sort of network transport (you can run a GUI application
>> on a Wayland display by using Xwayland, but that's sort of by-the-by).
> 
> Is this really true?  And yet, they expect everyone to switch from X to
> Wayland, when this is true?

I believe that waypipe¹ allows remote display with Wayland.

$ apt-cache search waypipe
waypipe - Network transparency with Wayland

I have not used it, as very few of my systems run Wayland at
this point.

¹ https://gitlab.freedesktop.org/mstoeckl/waypipe

-- 
Todd


signature.asc
Description: PGP signature

Re: Correct way to stop fetchmail

[Jerome BENOIT]

Hi, have you try

/etc/init.d/fetchmail stop

?

Cheers,
Jerome

On 28/12/2024 17:21, Roger Price wrote:

I would like to stop fetchmail on Debian 11.  I tried command systemctl stop 
fetchmail and system status fetchmail reported

  ● fetchmail.service - LSB: init-Script for system wide fetchmail daemon
  Loaded: loaded (/etc/init.d/fetchmail; generated)
  Active: inactive (dead) since Sat 2024-12-28 10:37:58 CET; 6h ago
    Docs: man:systemd-sysv-generator(8)
     Process: 1250 ExecStart=/etc/init.d/fetchmail start (code=exited, 
status=0/SUCCESS)
     Process: 1199929 ExecStop=/etc/init.d/fetchmail stop (code=exited, 
status=0/SUCCESS)
   Tasks: 1 (limit: 57750)
  Memory: 1.9M
     CPU: 6min 33.392s
  CGroup: /system.slice/fetchmail.service
  └─1264 /usr/bin/fetchmail -f /etc/fetchmarc
     --pidfile /var/run/fetchmail/fetchmail.pid --syslog
  ...
  Dec 28 17:06:54 titan fetchmail[1264]: reading message
     
mail...@rogerprice.org@mail.gandi.net:1 of 1 (9321 octets) flushed

It's dead, but still fetching mail! I tried

  root@titan ~ /etc/init.d/fetchmail --quit
  Not starting fetchmail daemon, disabled via /etc/default/fetchmail.

but fetchmail goes on fetching mail.  What is the correct way of stopping 
fetchmail?

Roger

Re: Correct way to stop fetchmail

[Roger Price]

On Sat, 28 Dec 2024, Greg Wooledge wrote:


On Sat, Dec 28, 2024 at 11:36:44 -0700, Charles Curley wrote:

On Sat, 28 Dec 2024 17:21:08 +0100 (CET)
Roger Price  wrote:


but fetchmail goes on fetching mail.  What is the correct way of
stopping fetchmail?


Perhaps there is a fetchmail.timer unit that is starting up
fetchmail.service from time to time?

Perhaps there's a cron job that starts it.


Debian 11 dpkg -L fetchmail does not show any timer unit or other cron device 
that might keep a "dead" service alive.



systemctl status fetchmail.timer


No such unit in Debian 11.


You might also look at /usr/share/doc/fetchmail/README.systemd


I found nothing that keeps officially "dead" services alive.


 root@titan ~ ps -ef | grep fetch
 fetchma+1264   1  0 Sep25 ?00:06:33 /usr/bin/fetchmail -f
  /etc/fetchmailrc --pidfile /var/run/fetchmail/fetchmail.pid --syslog
 root@titan ~ kill 1264



That process had been running for three months.(*)  It wasn't "dead for 6
hours" at the moment you ran those commands.


The process was alive for three months until I typed systemctl stop fetchmail. 6 
hours later I typed systemctl status fetchmail and systemd told me that the 
still running process had been "dead" for 6 hours.


# (*) Unless of course your system clock is massively wrong.

:-)  I rely on the Network Time Protocol (NTP).  An excellent service.

But the problem is solved.  Command kill came to the rescue.

Roger

Re: How to update fonts

[Charles Curley]

On Sat, 28 Dec 2024 19:28:47 +
Alain D D Williams  wrote:

> I occasionally see Unicode characters that do not do not display
> properly. Eg: 메리 크리스마스 (for the curious: this says Happy
> Christmas in Korean).

Well, I don't have that font installed, so I see little boxes with the
Unicode values in them, called tofu. So I grabbed one and searched in
the character map accessory (package gucharmap) for it. That told me it
was a Hangul symbol. A simple search gave me several Hangul font
packages one might try:

charles@hawk:~$ apt-cache search hangul | grep font
fonts-alee - free Hangul TrueType fonts
fonts-lexi-gulim - Lexi Gulim Korean font
fonts-lexi-saebom - Lexi Saebom Korean font
fonts-wqy-microhei - Sans-serif style CJK font derived from Droid
xfonts-efont-unicode - /efont/ Unicode fonts for X which cover various scripts
xfonts-efont-unicode-ib - /efont/ Unicode fonts for X (italic and bold)
charles@hawk:~$ 


-- 
Does anybody read signatures any more?

https://charlescurley.com
https://charlescurley.com/blog/

Re: synaptic workalike that WILL run on sudo with wayland.

[gene heskett]

On 12/28/24 15:49, Andy Smith wrote:

Hi,

On Sat, Dec 28, 2024 at 01:06:11PM -0500, gene heskett wrote:

On 12/28/24 08:24, Andy Smith wrote:

On Sat, Dec 28, 2024 at 12:53:41PM +, Darac Marjal wrote:

Synaptic is a GUI frontend to apt. It's not really much different to
aptitude, except that it's maybe a bit more mouse-friendly.

That much I know. I am wondering what its killer feature is that Gene
wants to work so hard to make it run on a headless single board computer
that one would think to be wholly unsuitable for running GUI apps.

What would have been useful for you to tell us at this point or in any
of your other replies:

- How exactly you are trying to run synaptic on these Banana Pi
   computers, e.g. with display and keyboard/mouse directly attached and
   you logged in through the Mate desktop environment, *or* over the
   network by SSH in a terminal from another computer that is running the
   Mate desktop environment.

What you instead chose to tell us:

which from this below indicates I do it both ways.

armbian in the full desktop version runs noticeably slower than it runs on
amd64 stuff, but its more than fast enough to run a full screen gui for a 3d
printer.  I'll gladly tolerate a just noticeable lag that never crashes,
uptimes from kernel update to kernel update, with klipper/moonraker/fluidd
feeding nginx, broadcasting to any web browser in the house that is watching
aliasname:80, in exchange for an 18 watt power draw (including its 24" AOL
monitor) while sitting idle between jobs. I use firefox to localhost:80 on
the local machine actually running the printer at speeds around 10x what it
could do OOTB 4 years ago when it was new.

FWIW, I probably have another $2000 in better hdwe., much of which I
designed in OpenSCAD. I may ask what seems to be off the wall questions
simply hoping to glean additional info because I enjoy plowing new ground
and making it work better.  Part of the upgrade is closed loop
stepper/servos, because they have at least a 10x improvement in the accuracy
of motor control. The error at the instant determines motor current, running
at 7x speeds, average motor power is 30% of the burn your hand temps the
usual setup does, it can even reverse a motor that overshoots, putting it
back where it belongs using around 4x the instant power if it has to. The
latest addition is a motorized nozzle cleaner, operated as part of the
startup gcode.

I've got to get my web page up again, I lost it when those two seagate 2T
drives went offline at 2 weeks accumulated runtime as I was trying make
bookworm run now 18 months ago. It runs when it gets around to it, opening a
local file is an automatic lockup for 30 seconds. And no one has identified
that problem. Around 30 fresh installs with the only usb plugged in being
the wireless button for logitek keyboard/mouse but I still get orca and
brltty yelling every keystroke at me thru my speakers. Yet when I question
the broken bookworm installer, I catch hell from the powers that be here.
Total denial that the problem exists has been the universal response.

Thanks,
Andy


Cheers, Gene Heskett, CET.
--
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis

Re: double-boot laptop with Debian Trixie

[David Wright]

On Sun 29 Dec 2024 at 01:20:17 (+0100), Rafał Lichwała wrote:
> 
> I plan to switch completely from Windows 11 to Debian on my laptop,
> but as a first step I'd like to configure double-boot setup to give it
> a try, look around and check if all hardware works fine under Debian.
> 
> Current state:
> Dell laptop (one NVMe disk only) with Windows 11 onboard.
> Windows starts in UEFI mode (Secure Boot ON).
> Already resized main partition to make space for Debian.
> So currently there are the following partitions on disk:
> 
> 100MB, FAT32, GPT (EFI System Partition)
> 16MB, Other, GPT (Reserved Partition)
> 300GB, NTFS, GPT (Data Partition) (Main C: disk visible under Win11)
> 168GB, free space, ready to have ~150GB for Debian and 16GB for swap
> 880MB, NTFS, GPT (Recovery Partition)

> Third try:
> Boot in Legacy Mode (so Secure Boot is OFF).
> Debian Trixie can see my laptop disk properly, so installation is
> successful, but...
> It cannot see my Windows 11 installation when trying to
> install/configure/prepare GRUB.

It doesn't need to. Switch to UEFI mode, and the ESP will
boot into Windows.

> At this step I decided to NOT install GRUB on my main disk to NOT
> loose access to my Windows 11 system.

If you booted in Legacy Mode, then Grub will install itself
into the protective MBR. If it then asks about installing
“an extra copy of the EFI version of the GRUB boot loader
to a fallback location, the "removable media path"”, say No.

> (but Trixie says that it's possible to restore access by manual
> configuration of GRUB later on - I'm not sure...)

Yes, that's on the same screen as I just mentioned. But you
say No, so you don't touch that fallback in the ESP.¹

However, when you install Grub in Legacy Mode on a GPT disk,
you ought to make a small partition with code EF02, a BIOS
Boot Partition. 4MB will do, and doesn't mess up alignments.
Grub puts its core.img there. It's not essential: Grub can
put core.img into your linux partition using blocklists, but
this technique is fragile. It would do for the time being
while you're trying out Debian in the presence of Windows.

When you eventually wipe Windows, you'll be installing in,
or converting to, UEFI mode anyway.

> Main goal:
> To have both current Windows 11 and new Debian Trixie available in
> GRUB and properly bootable.

Both would be bootable, with the current setting of UEFI or
Legacy Mode determining which one boots this time around.

BTW I'd recommend bookworm to start with, as others have.

¹ I've had that screen come up on machines that predate EFI,
  where the first sentence is obviously nonsense.

Cheers,
David.

Re: synaptic workalike that WILL run on sudo with wayland.

[Chris Green]

gene heskett  wrote:
> 
> On 12/28/24 08:24, Andy Smith wrote:
> > Hi,
> >
> > On Sat, Dec 28, 2024 at 12:53:41PM +, Darac Marjal wrote:
> >> On 28/12/2024 10:06, Andy Smith wrote:
> >>> On Fri, Dec 27, 2024 at 08:26:26PM -0500, gene heskett wrote:
>  When is that going to happen? Trying to maintain debian-arm stuff blind 
>  is a
>  pita. I want to SEE whats available.
> >>> I have never in my life felt the need to use synaptic. What am I
> >>> missing out on?
> >> Synaptic is a GUI frontend to apt. It's not really much different to
> >> aptitude, except that it's maybe a bit more mouse-friendly.
> > That much I know. I am wondering what its killer feature is that Gene
> > wants to work so hard to make it run on a headless single board computer
> > that one would think to be wholly unsuitable for running GUI apps.
> 
> armbian in the full desktop version runs noticeably slower than it runs 
> on amd64 stuff, but its more than fast enough to run a full screen gui 
> for a 3d printer. 

Surely that (the speed) is down to how fast the particular ARM or X86
system is.  There's fast ARM systems and slow X86 ones (though I
woiuld guess that the fastest X86 ones are faster than the fastest ARM
ones).

-- 
Chris Green
·

Re: Correct way to stop fetchmail

[Greg Wooledge]

On Sat, Dec 28, 2024 at 22:59:19 +0100, Nicolas George wrote:
> Roger Price (12024-12-28):
> >  ● fetchmail.service - LSB: init-Script for system wide fetchmail daemon
> >  Loaded: loaded (/etc/init.d/fetchmail; generated)
> >  Active: inactive (dead) since Sat 2024-12-28 10:37:58 CET; 6h ago
> >Docs: man:systemd-sysv-generator(8)
> > Process: 1250 ExecStart=/etc/init.d/fetchmail start (code=exited, 
> > status=0/SUCCESS)
> > Process: 1199929 ExecStop=/etc/init.d/fetchmail stop (code=exited, 
> > status=0/SUCCESS)
> >   Tasks: 1 (limit: 57750)
> >  Memory: 1.9M
> > CPU: 6min 33.392s
> >  CGroup: /system.slice/fetchmail.service
> >  └─1264 /usr/bin/fetchmail -f /etc/fetchmarc
> > --pidfile /var/run/fetchmail/fetchmail.pid --syslog
> >  ...
> >  Dec 28 17:06:54 titan fetchmail[1264]: reading message
> > 
> > mail...@rogerprice.org@mail.gandi.net:1 of 1 (9321 octets) flushed
> > 
> > It's dead
> 
> It is “dead” but not dead, since it still has a process. Your problem is
> there, you do not need to look elsewhere: systemctl stop did not do its
> job properly. You need to find out why.

If you look at the ExecStart= there, you can see that this isn't a
proper systemd unit.  It's just an auto-generated wrapper around the
init.d scripts.

If an init.d script is broken and doesn't stop the service that it's
supposed to stop, well, that's kinda why people stopped using sysv init.
It's really terrible.

I believe Roger said he's on Debian 11(?).  The file list for the Deb12
package shows both an init.d script and a systemd unit file.


I don't know whether one of them is just a stub that acts as a wrapper
for the other, or whether they're both proper, full-fledged management
files for their respective init systems.

So, this might be a "bug already fixed in stable" kind of situation.

qt6-svg-plugins should be a hard dependency from kde-full or task-kde-desktop

[Christopher Andrews]

I am unsure exactly what package to file this bug against, so per the 
recommendation I am sending to this list. I suspect this should be filed 
against libqt6gui6.


I run debian/unstable, and my system somehow got into a state where both 
kde-full and task-kde-desktop is installed, but all the SVG icons on the 
desktop were missing. I searched for packages with svg in the name and 
found that "qt6-svg-plugins" was missing, installing it brang back all 
the SVG icons. I retested, removing this package breaks the icons, 
bringing it back restores them.


I had assumed that the transition to plasma6 was an issue, but after 
weeks of missing icons I decided to hunt the problem down.


I suggest switching "qt6-svg-plugins" from a recommends to a depends in 
libqt6gui6 (or any more appropriate package that is eventually depended 
by the metapackage) as visible desktop icons probably should be part of 
the "full" install, and it's currently possible to simply uninstall it.

Re: synaptic workalike that WILL run on sudo with wayland.

[Andy Smith]

Hi,

On Sat, Dec 28, 2024 at 01:06:11PM -0500, gene heskett wrote:
> On 12/28/24 08:24, Andy Smith wrote:
> > On Sat, Dec 28, 2024 at 12:53:41PM +, Darac Marjal wrote:
> > > Synaptic is a GUI frontend to apt. It's not really much different to
> > > aptitude, except that it's maybe a bit more mouse-friendly.
> > That much I know. I am wondering what its killer feature is that Gene
> > wants to work so hard to make it run on a headless single board computer
> > that one would think to be wholly unsuitable for running GUI apps.

What would have been useful for you to tell us at this point or in any
of your other replies:

- How exactly you are trying to run synaptic on these Banana Pi
  computers, e.g. with display and keyboard/mouse directly attached and
  you logged in through the Mate desktop environment, *or* over the
  network by SSH in a terminal from another computer that is running the
  Mate desktop environment.

What you instead chose to tell us:

> armbian in the full desktop version runs noticeably slower than it runs on
> amd64 stuff, but its more than fast enough to run a full screen gui for a 3d
> printer.  I'll gladly tolerate a just noticeable lag that never crashes,
> uptimes from kernel update to kernel update, with klipper/moonraker/fluidd
> feeding nginx, broadcasting to any web browser in the house that is watching
> aliasname:80, in exchange for an 18 watt power draw (including its 24" AOL
> monitor) while sitting idle between jobs. I use firefox to localhost:80 on
> the local machine actually running the printer at speeds around 10x what it
> could do OOTB 4 years ago when it was new.
> 
> FWIW, I probably have another $2000 in better hdwe., much of which I
> designed in OpenSCAD. I may ask what seems to be off the wall questions
> simply hoping to glean additional info because I enjoy plowing new ground
> and making it work better.  Part of the upgrade is closed loop
> stepper/servos, because they have at least a 10x improvement in the accuracy
> of motor control. The error at the instant determines motor current, running
> at 7x speeds, average motor power is 30% of the burn your hand temps the
> usual setup does, it can even reverse a motor that overshoots, putting it
> back where it belongs using around 4x the instant power if it has to. The
> latest addition is a motorized nozzle cleaner, operated as part of the
> startup gcode.
> 
> I've got to get my web page up again, I lost it when those two seagate 2T
> drives went offline at 2 weeks accumulated runtime as I was trying make
> bookworm run now 18 months ago. It runs when it gets around to it, opening a
> local file is an automatic lockup for 30 seconds. And no one has identified
> that problem. Around 30 fresh installs with the only usb plugged in being
> the wireless button for logitek keyboard/mouse but I still get orca and
> brltty yelling every keystroke at me thru my speakers. Yet when I question
> the broken bookworm installer, I catch hell from the powers that be here.
> Total denial that the problem exists has been the universal response.

Thanks,
Andy

-- 
https://bitfolk.com/ -- No-nonsense VPS hosting

Re: synaptic workalike that WILL run on sudo with wayland.

[gene heskett]

On 12/28/24 16:41, Andrew M.A. Cater wrote:

On Sat, Dec 28, 2024 at 01:06:11PM -0500, gene heskett wrote:

armbian in the full desktop version runs noticeably slower than it runs on
amd64 stuff, but its more than fast enough to run a full screen gui for a 3d
printer.  I'll gladly tolerate a just noticeable lag that never crashes,
uptimes from kernel update to kernel update, with klipper/moonraker/fluidd
feeding nginx, broadcasting to any web browser in the house that is watching
aliasname:80, in exchange for an 18 watt power draw (including its 24" AOL
monitor) while sitting idle between jobs. I use firefox to localhost:80 on
the local machine actually running the printer at speeds around 10x what it
could do OOTB 4 years ago when it was new.


Armbian is *NOT* Debian  They do things differently there. Some of their
images are based on Ubuntu, others on Debian. They usually use the BSP -
Board Support Package - from the hardware vendor but that could be using any
kernel. We've gone through all of this multiple times, Gene: the reason
no one can support you on Armbian is because it's basically a different
and unknown quantity. It may be sufficently similar that we can extrapolate
but it's like taking a Volkswagen into a Toyota dealer and anticipating they
can fix it immediately with no prior knowledge. It might be like taking
an EV into your friendly mechanic who has never seen one - who knows how
long the learning curve is for all of us. That's before we get into the fact
that most of us have never seen a BananaPi 5 and never will.
Then you are missing out on the many things a pi clone can do on 5% of 
an amd64's power budget. Granted the amd64 can do it 20x faster, 
ignoring that the pi clone is fast enough. Debian seems to treat the 
arm64's a toy.



That is called putting the tech to work, more than earning its keep.

I've got to get my web page up again, I lost it when those two seagate 2T
drives went offline at 2 weeks accumulated runtime as I was trying make
bookworm run now 18 months ago. It runs when it gets around to it, opening a
local file is an automatic lockup for 30 seconds. And no one has identified
that problem. Around 30 fresh installs with the only usb plugged in being
the wireless button for logitek keyboard/mouse but I still get orca and
brltty yelling every keystroke at me thru my speakers. Yet when I question
the broken bookworm installer, I catch hell from the powers that be here.
Total denial that the problem exists has been the universal response.


The lock ups are indeterminate because you can't ever tell us what you've done

bs, I have described exactly what I've done but you treat it as TLDR.

to get into this state. At this point, bringing up orca and brltty is once
again old news and not relevant. Ten days or two weeks ago, you promised
that this would be the end of this because you'd go and finally do a clean
install as install 31. Please do so - simplify everything, unmount your RAID
until you've done a basic install and come back when you can show history
of what you've done. I don't deny that you have problems but the vagueness
of response over a period of years doesn't help you plead your case.
And the bookworm install isn't irremediably broken - thousands of installs
by other people suggest that the problem may be yours and yours alone.


I'll not argue  that point but I can't cross a wireless keyboard/mouse 
button receiver button as being responsible for this when I've 
repeatedly been told its because the installer found a serial adaptor 
whose feed cable to an x10  cm11a is laying on the floor unplugged 
during the many installs. it isn't wireless and its unplugged. If that 
is not clear enough I don't know how else to explain it. To strip this 
machine down to just one drive, reinstall yet again, then remount my 
/home drive over what the installer makes is a major undertaking as 
there's currently 2 other drive controllers to be removed or disabled, 
installed alongside the 6 sata-ii ports this board has.  I'm not willing 
to take the chance the installer might decide to format everything else 
it finds. I've had that actually happen, loosing a couple months work I 
had to re-invent. Back around stretch or jessie time,  No thanks. The 
installer should NOT decide to install stuff I have purposely skipped 
during the install, but it has installed and made live orca and brltty 
around 30 times. Most of those 30 installs were because by the time I 
had stuffed a hot potato into orcas mouth by removing the exec bits from 
the perms, it was not possible to reboot, it got stuck waiting for orca 
the come live, so a reboot was done by reinstalling.


With nothing plugged into a usb port but the logitek button.  I don't 
own a wired mouse except the 37 yo 1200 baud serial mouse on my now 
deceased from dried out caps color computer 3.  And there is not even a 
compatible db9 on this computer. Time marches on. The person running the 
installer should have the final sa

double-boot laptop with Debian Trixie

[Rafał Lichwała]

Hi,

I plan to switch completely from Windows 11 to Debian on my laptop, but 
as a first step I'd like to configure double-boot setup to give it a 
try, look around and check if all hardware works fine under Debian.


Current state:
Dell laptop (one NVMe disk only) with Windows 11 onboard.
Windows starts in UEFI mode (Secure Boot ON).
Already resized main partition to make space for Debian.
So currently there are the following partitions on disk:

100MB, FAT32, GPT (EFI System Partition)
16MB, Other, GPT (Reserved Partition)
300GB, NTFS, GPT (Data Partition) (Main C: disk visible under Win11)
168GB, free space, ready to have ~150GB for Debian and 16GB for swap
880MB, NTFS, GPT (Recovery Partition)

First try:
UEFI Boot (Secure Boot ON), Debian Trixie, Expert mode installation
During installation Debian does NOT see my disk at all (pendrive with 
Debian Trixie is visible only)
Also during installation Debian says that "it seems there is another 
system installed in Legacy Mode, am I  sure to force UEFI mode?" which 
is not correct - Win11 is installed in UEFI mode (when I try to boot it 
in Legacy Mode I get Windows Blue Screen Of Death)


Second try:
Secure Boot set to OFF.
Nothing changed during Trixie installation, so no device to install on.

Third try:
Boot in Legacy Mode (so Secure Boot is OFF).
Debian Trixie can see my laptop disk properly, so installation is 
successful, but...
It cannot see my Windows 11 installation when trying to 
install/configure/prepare GRUB.
At this step I decided to NOT install GRUB on my main disk to NOT loose 
access to my Windows 11 system.
(but Trixie says that it's possible to restore access by manual 
configuration of GRUB later on - I'm not sure...)


Main goal:
To have both current Windows 11 and new Debian Trixie available in GRUB 
and properly bootable.


How to achieve this? :-)

Thank you,
Best Regards,
Rafal

Re: synaptic workalike that WILL run on sudo with wayland.

[Andrew M.A. Cater]

On Sat, Dec 28, 2024 at 01:06:11PM -0500, gene heskett wrote:
> 
> armbian in the full desktop version runs noticeably slower than it runs on
> amd64 stuff, but its more than fast enough to run a full screen gui for a 3d
> printer.  I'll gladly tolerate a just noticeable lag that never crashes,
> uptimes from kernel update to kernel update, with klipper/moonraker/fluidd
> feeding nginx, broadcasting to any web browser in the house that is watching
> aliasname:80, in exchange for an 18 watt power draw (including its 24" AOL
> monitor) while sitting idle between jobs. I use firefox to localhost:80 on
> the local machine actually running the printer at speeds around 10x what it
> could do OOTB 4 years ago when it was new.
> 

Armbian is *NOT* Debian  They do things differently there. Some of their
images are based on Ubuntu, others on Debian. They usually use the BSP -
Board Support Package - from the hardware vendor but that could be using any
kernel. We've gone through all of this multiple times, Gene: the reason
no one can support you on Armbian is because it's basically a different
and unknown quantity. It may be sufficently similar that we can extrapolate
but it's like taking a Volkswagen into a Toyota dealer and anticipating they
can fix it immediately with no prior knowledge. It might be like taking 
an EV into your friendly mechanic who has never seen one - who knows how
long the learning curve is for all of us. That's before we get into the fact
that most of us have never seen a BananaPi 5 and never will.


> 
> I've got to get my web page up again, I lost it when those two seagate 2T
> drives went offline at 2 weeks accumulated runtime as I was trying make
> bookworm run now 18 months ago. It runs when it gets around to it, opening a
> local file is an automatic lockup for 30 seconds. And no one has identified
> that problem. Around 30 fresh installs with the only usb plugged in being
> the wireless button for logitek keyboard/mouse but I still get orca and
> brltty yelling every keystroke at me thru my speakers. Yet when I question
> the broken bookworm installer, I catch hell from the powers that be here.
> Total denial that the problem exists has been the universal response.
> 

The lock ups are indeterminate because you can't ever tell us what you've done
to get into this state. At this point, bringing up orca and brltty is once
again old news and not relevant. Ten days or two weeks ago, you promised
that this would be the end of this because you'd go and finally do a clean
install as install 31. Please do so - simplify everything, unmount your RAID
until you've done a basic install and come back when you can show history
of what you've done. I don't deny that you have problems but the vagueness
of response over a period of years doesn't help you plead your case.
And the bookworm install isn't irremediably broken - thousands of installs
by other people suggest that the problem may be yours and yours alone.
> > 



All the very best, as ever,

Andrew Cater
(amaca...@debian.org)

> Cheers, Gene Heskett, CET.
> -- 
> "There are four boxes to be used in defense of liberty:
>  soap, ballot, jury, and ammo. Please use in that order."
> -Ed Howdershelt (Author, 1940)
> If we desire respect for the law, we must first make the law respectable.
>  - Louis D. Brandeis
> 

Re: Correct way to stop fetchmail

[Nicolas George]

Roger Price (12024-12-28):
>  ● fetchmail.service - LSB: init-Script for system wide fetchmail daemon
>  Loaded: loaded (/etc/init.d/fetchmail; generated)
>  Active: inactive (dead) since Sat 2024-12-28 10:37:58 CET; 6h ago
>Docs: man:systemd-sysv-generator(8)
> Process: 1250 ExecStart=/etc/init.d/fetchmail start (code=exited, 
> status=0/SUCCESS)
> Process: 1199929 ExecStop=/etc/init.d/fetchmail stop (code=exited, 
> status=0/SUCCESS)
>   Tasks: 1 (limit: 57750)
>  Memory: 1.9M
> CPU: 6min 33.392s
>  CGroup: /system.slice/fetchmail.service
>  └─1264 /usr/bin/fetchmail -f /etc/fetchmarc
> --pidfile /var/run/fetchmail/fetchmail.pid --syslog
>  ...
>  Dec 28 17:06:54 titan fetchmail[1264]: reading message
> 
> mail...@rogerprice.org@mail.gandi.net:1 of 1 (9321 octets) flushed
> 
> It's dead

It is “dead” but not dead, since it still has a process. Your problem is
there, you do not need to look elsewhere: systemctl stop did not do its
job properly. You need to find out why.

-- 
  Nicolas George

Re: double-boot laptop with Debian Trixie

[David Christensen]

On 12/28/24 16:20, Rafał Lichwała wrote:

Hi,

I plan to switch completely from Windows 11 to Debian on my laptop, 



Make?  Dell, below.  It is better to state this information at the top 
of your message.



Model?


but 
as a first step I'd like to configure double-boot setup to give it a 
try, look around and check if all hardware works fine under Debian.



I recommend that you use Debian Live "Statble" to test hardware 
compatibility.:


https://www.debian.org/CD/live/



Current state:
Dell laptop (one NVMe disk only) with Windows 11 onboard.
Windows starts in UEFI mode (Secure Boot ON).
Already resized main partition to make space for Debian.
So currently there are the following partitions on disk:

100MB, FAT32, GPT (EFI System Partition)
16MB, Other, GPT (Reserved Partition)
300GB, NTFS, GPT (Data Partition) (Main C: disk visible under Win11)
168GB, free space, ready to have ~150GB for Debian and 16GB for swap
880MB, NTFS, GPT (Recovery Partition)



Manual transcription is error prone.  It is better to cut and paste a 
console session showing the exact prompt, command entered, and output 
displayed.




First try:
UEFI Boot (Secure Boot ON), Debian Trixie, Expert mode installation



Trixie is still in testing:

https://www.debian.org/releases/trixie/


It is better to use Stable (or Old Stable).


During installation Debian does NOT see my disk at all (pendrive with 
Debian Trixie is visible only)



On my Dell machines with NVMe drives, I must change the Setup settings 
for Debian to see NVMe drives:


-> Settings
   -> System Configuration
   -> SATA Operation -> AHCI


Also during installation Debian says that "it seems there is another 
system installed in Legacy Mode, am I  sure to force UEFI mode?" which 
is not correct - Win11 is installed in UEFI mode (when I try to boot it 
in Legacy Mode I get Windows Blue Screen Of Death)


Second try:
Secure Boot set to OFF.
Nothing changed during Trixie installation, so no device to install on.

Third try:
Boot in Legacy Mode (so Secure Boot is OFF).
Debian Trixie can see my laptop disk properly, so installation is 
successful, but...
It cannot see my Windows 11 installation when trying to install/ 
configure/prepare GRUB.
At this step I decided to NOT install GRUB on my main disk to NOT loose 
access to my Windows 11 system.
(but Trixie says that it's possible to restore access by manual 
configuration of GRUB later on - I'm not sure...)


Main goal:
To have both current Windows 11 and new Debian Trixie available in GRUB 
and properly bootable.


How to achieve this? :-)

Thank you,
Best Regards,
Rafal



It is best to take an image of the system drive before making 
significant changes.  I recommend that you do so now, before making 
further changes.



I recommend that you get Windows 11 working again, use Windows Disk 
Management remove the partitions created by d-i, create a partition and 
filesystem for virtual machines in the empty space, install your 
hypervisor of choice, create a VM, and install Debian Stable into the VM.



David

Re: What's best way to handle HTML emails in Mutt?

[Christoph Brinkhaus]

Am Fri, Dec 27, 2024 at 11:41:39PM -0500 schrieb hobie of RMN:
> For 20 years I've enjoyed Mutt as my primary mail reader.  I used its
> 'bounce' feature to deal with HTML mails, sending them to a webmail
> program on a different server.  But now the jellyfish anti-spam daemon
> won't allow me to do that; it rejects my bounced emails because "From:"
> and "(envelope-from)" are not the same.
> 
> What's the best way to handle this?  Switch to Thunderbird or claws-mail?
> 
May be you can use the send-hook for that purpose.
See below an example of my muttrc. Usually I use my t-online account.
With the send-hook the "From:" is changed in case the receiver has a
gmx address.

send-hook . 'unmy_hdr From:'
send-hook '~t ^.*@gmx\.net$' 'my_hdr From: Christoph Brinkhaus 
'

For your issue -t might not help, additionally I do not know if
envelope-from can be changed by a send-hook. But the example should show
that it is not pure black magic. May be it is a starting point for you.

Kind regards,
Christoph
-- 
Ist die Katze gesund
schmeckt sie dem Hund.


signature.asc
Description: PGP signature

Re: synaptic workalike that WILL run on sudo with wayland.

[Darac Marjal]

On 28/12/2024 10:06, Andy Smith wrote:

Hi,

On Fri, Dec 27, 2024 at 08:26:26PM -0500, gene heskett wrote:

When is that going to happen? Trying to maintain debian-arm stuff blind is a
pita. I want to SEE whats available.

I have never in my life felt the need to use synaptic. What am I
missing out on?
Synaptic is a GUI frontend to apt. It's not really much different to 
aptitude, except that it's maybe a bit more mouse-friendly.

I just use "apt" from a terminal. apticron emails me which updates are
available.

Running headless servers but still wanting a GUI app on each and every
one of them to manage updates? Doesn't sound like the Linux I know!


You can't run a GUI on a headless server. That's the whole point of it 
being headless. You can run X clients on the server and remotely display 
that on a workstation (i.e. something _with_ a head), but AFAIK the 
wayland protocol doesn't support any sort of network transport (you can 
run a GUI application on a Wayland display by using Xwayland, but that's 
sort of by-the-by).


OK. /Technically/ you can run a GUI on the headless server. You can use 
things like Xvfb to render the display on the headless server, but 
you're starting to get into "how many ways to skin a cat". Do you want 
to send pixels over the network, or do you want to send a stream of 
drawing instructions?





Thanks,
Andy



OpenPGP_signature.asc
Description: OpenPGP digital signature

Re: Trixie and Nvidia proprietary drivers

[George at Clug]

On Saturday, 28-12-2024 at 21:36 George at Clug wrote:
> ​Hi,
> 
> 
> After installing updates this morning, my Debian Trixie, KDE X11
> installation hangs while loading Trixie with Nvidia proprietary
> drivers as packaged with Debian Trixie (the lockup is before the logon
> screen appears).
> 
> 
> My Nvidia GPU is a RTX 2070. Intel i5 processor.
> 
> 
> I did a reboot and logged in maintenance mode, removed the Nvidia
> drivers (" # apt purge nvidia* ") which allowed me to log in with
> Nouveau drivers. 
> 
> 
> 
> When I reinstalled the proprietary Nvidia drivers I had the same issue
> of the boot process hanging, and had to go back to Nouveau drivers. 
> Which I am currently using.
> 
> 
> I am wondering if anyone else has experience this issue?
> 
> 
> I expect if I wait long enough someone will have resolved the issue.
> 
> 
> I am so curious if and when Debian Trixie will be packaged with the
> Nvidia stable version of the NVIDIA 565.77 drivers?


https://tracker.debian.org/pkg/nvidia-open-gpu-kernel-modules
nvidia-open-gpu-kernel-modules
 A new upstream version is available: 565.77 high
A new upstream version 565.77 is available, you should consider packaging it. 


>   I am looking
> forward to this, also hoping they are introduced in Debian Bookworm,
> now that the drivers are 'stable'. My guess is that KDE Plasma 6 will
> work well with Wayland using these drivers.


> 
> 
> 
> https://9to5linux.com/nvidia-565-linux-graphics-driver-is-now-available-for-download-heres-whats-new
> NVIDIA released today NVIDIA 565.77 as the first stable version of the
> NVIDIA 565 graphics driver series for Linux, FreeBSD, and Solaris
> systems with various improvements and multiple bug fixes.
> 
> 
> George.
>