Lawrence D'Oliveiro <ldo@nz.invalid> wrote:
> I recently discovered this French phrase, ?Syst?me D?, where the D is
> supposed to stand for ?debrouillardise?. It basically refers to an ability
> to improvise and adapt your way around challenges, in whatever situation
> you encounter.

The phrase is also used by English-speaking economists, who usually render
it as "System D", which refers to the sort of unregulated, informal
economy you might find in a third-world country, where everything is
makeshift and improvised. This contrasts with Systemd, which has more
in common with communism. It demands that everything be remade the
Systemd way, and has forced many poorly-reinvented wheels on us in
the process (even DNS is still broken, after all these years, thanks
to systemd-resolved). Even sshd had to be made to depend on it, even
though the upstream refused to support it, resulting in the now-infamous
XZ Utils backdoor being relevant to sshd through a transient dependency
on libxz by libsystemd. On non-Systemd operating systems, the XZ Utils
backdoor did not affect sshd.

On Wed, 15 May 2024 09:08:31 -0000 (UTC), Sebastian obviously hasn’t read
the biggest myths about systemd <http://0pointer.de/blog/projects/the-
biggest-myths.html>:

> This contrasts with Systemd, which has more in common with communism. It
> demands that everything be remade the Systemd way ...

It doesn’t demand anything. It’s Open Source, which means it respects your
freedoms. Also, like all good system software, it is designed to offer
“mechanism, not policy”. It provides the mechanisms, which you can
configure to implement whatever policies you like.

> ... and has forced many poorly-reinvented wheels on
> us in the process ...

It is made up of modular pieces. Outside of the core (init, journald,
udevd), everything else is optional.

> ... Even sshd had to be made to depend on it ...

That was purely a distro choice, to take advantage of the more robust
service-management features that systemd offers.

On Fri, 17 May 2024 21:46:19 -0000 (UTC), Lawrence D'Oliveiro wrote:

>
> That was purely a distro choice, to take advantage of the more robust
> service-management features that systemd offers.
>

Another lackey, ignoramus statement.

GNU/Linux, at least on the desktop workstation, does NOT require
"services," let alone "service management."

I operate a totally potent desktop workstation and I have NO
"services." None. Zip. Nada.

Thus your lackey contention is full of fucking shit.

Sytemd is an abomination and should have been strangled
in the cradle. But dupes like you have kept it alive
to wreak its destruction.

Lawrence D'Oliveiro wrote this copyrighted missive and expects royalties:

> On Wed, 15 May 2024 09:08:31 -0000 (UTC), Sebastian obviously hasn’t read
> the biggest myths about systemd <http://0pointer.de/blog/projects/the-
> biggest-myths.html>:
>
>> This contrasts with Systemd, which has more in common with communism. It
>> demands that everything be remade the Systemd way ...
>
> It doesn’t demand anything. It’s Open Source, which means it respects your
> freedoms. Also, like all good system software, it is designed to offer
> “mechanism, not policy”. It provides the mechanisms, which you can
> configure to implement whatever policies you like.
>
>> ... and has forced many poorly-reinvented wheels on
>> us in the process ...
>
> It is made up of modular pieces. Outside of the core (init, journald,
> udevd), everything else is optional.
>
>> ... Even sshd had to be made to depend on it ...
>
> That was purely a distro choice, to take advantage of the more robust
> service-management features that systemd offers.

These morons think every new system is "communism".

--
Go not to the elves for counsel, for they will say both yes and no.
-- J.R.R. Tolkien

Lawrence D'Oliveiro <ldo@nz.invalid> wrote:
> On Wed, 15 May 2024 09:08:31 -0000 (UTC), Sebastian obviously hasn?t read
> the biggest myths about systemd <http://0pointer.de/blog/projects/the-
> biggest-myths.html>:
>
>> This contrasts with Systemd, which has more in common with communism. It
>> demands that everything be remade the Systemd way ...
>
> It doesn?t demand anything. It?s Open Source, which means it respects your
> freedoms. Also, like all good system software, it is designed to offer
> ?mechanism, not policy?. It provides the mechanisms, which you can
> configure to implement whatever policies you like.

>
>> ... and has forced many poorly-reinvented wheels on
>> us in the process ...
>
> It is made up of modular pieces. Outside of the core (init, journald,
> udevd), everything else is optional.
>

Getting rid of non-core Systemd components is easier said than done on most
distros. You didn't list systemd-logind in the "core" components,
implying it's optional, but try setting up a Debian or Fedora system
without it. Even Devuan, whose reason for existence is to be a
Systemd-free Debian, had to use a fork of systemd-logind rather than
just handling sessions the pre-Systemd way, because so many things
have been made to depend on it.

Le 17-05-2024, Lawrence D'Oliveiro <ldo@nz.invalid> a écrit :
> On Wed, 15 May 2024 09:08:31 -0000 (UTC), Sebastian obviously hasn’t read
> the biggest myths about systemd <http://0pointer.de/blog/projects/the-
> biggest-myths.html>:
>
>> ... and has forced many poorly-reinvented wheels on
>> us in the process ...
>
> It is made up of modular pieces. Outside of the core (init, journald,
> udevd), everything else is optional.

Yes. For example, Guix use shepherd instead of systemd. But they kept logind.
Everything done by choice.

--
Si vous avez du temps à perdre :
https://scarpet42.gitlab.io

Le 17-05-2024, Farley Flud <ff@linux.rocks> a écrit :
> On Fri, 17 May 2024 21:46:19 -0000 (UTC), Lawrence D'Oliveiro wrote:
>
>>
>> That was purely a distro choice, to take advantage of the more robust
>> service-management features that systemd offers.
>>
>
> Another lackey, ignoramus statement.

Another great post where you are looking through a mirror.

> GNU/Linux, at least on the desktop workstation, does NOT require
> "services," let alone "service management."

You are confused. Once again. As your computer does nothing, it requires
nothing else than a compiler to be able to fail miserably to compile
your kernel.

But a useful workstation can be improved greatly with service
management. And systemd is very good at it.

> I operate a totally potent desktop workstation and I have NO
> "services." None. Zip. Nada.

Yes, of course your desktop is useless, no need of service for staring
at failing compilation logs endlessly.

> Thus your lackey contention is full of fucking shit.

It's the difference between looking stupidly a once computer and using
it efficiently.

> Sytemd is an abomination

There are issues with systemd. Granted. But there is nothing good enough
to replace it. As you are the best programmer in pretence in the world,
you just have one thing to do: write something that does everything that
systemd is good at without his issues. And everyone will switch to it.

As systemd was created by only one guy far less good at programming than
you, you should be able to do it. Finger in the nose. During the night.
And I wonder why you didn't did it yet. Maybe you are not as good as
your claims?

> and should have been strangled in the cradle. But dupes like you have
> kept it alive to wreak its destruction.

Do best and systemd will go in history books. Just brag as you only
know and it will stay on the computers. It's your choice.

--
Si vous avez du temps à perdre :
https://scarpet42.gitlab.io

On 19 May 2024 10:25:07 GMT, Stéphane CARPENTIER wrote:

>
> But a useful workstation can be improved greatly with service
> management.
>

How so?

The purpose of a workstation is to run software.

If my software requires a "service," I can manage it myself.

>
> And systemd is very good at it.
>

Anything that requires 3 million lines of code to do
its job cannot be described as "good."

>
> write something that does everything that
> systemd is good at without his issues.
>

I already have written an init/service management system.

It is customized for my needs only.

GNU/Linux makes this possible for me and for everyone else,
but systemd threatens to destroy the possibility for everyone.

On Sun, 19 May 2024 09:03:06 -0000 (UTC), Sebastian wrote:

> Even Devuan, whose reason for existence is to be a
> Systemd-free Debian, had to use a fork of systemd-logind rather than
> just handling sessions the pre-Systemd way, because so many things
> have been made to depend on it.
>

It will only get worse.

As I have indicated many times in the past, the goal of systemd,
as stated by that jackass Poettering himself, is to become the
one-and-only interface between the kernel and user space.

This goal is being facilitated by RedHat and freedesktop.org
and eventually there will be only a single "flavor" of GNU/Linux.

The systemd movement can be stopped but it would take strong
protests and most users are totally subservient distro lackeys.

Le 19-05-2024, Farley Flud <ff@linux.rocks> a écrit :
> On 19 May 2024 10:25:07 GMT, Stéphane CARPENTIER wrote:
>
>>
>> But a useful workstation can be improved greatly with service
>> management.
>>
>
> How so?
>
> The purpose of a workstation is to run software.

Not only and some softwares needs services. For the not only parts, I
want to be able to [un]plug easily a USB key/drive, a mouse, a printer, the
WiFi, well anything I didn't plugged before starting my computer. And
for the softwares, when they run in docker, the service need to be
started. A lot of things like the logs are very well managed thanks to a
service.

> If my software requires a "service," I can manage it myself.

I'm not that sure about that. And if you manage to start it
automatically at the boot time, you won't be able to see when, a service
crashes and need a restart.

For example I remember when a few weeks/months ago you had issues with
your ntpd service. You are the only one I know who were able to break
it. That tells a lot.

>> And systemd is very good at it.
>
> Anything that requires 3 million lines of code to do
> its job cannot be described as "good."

So you are publicly telling everyone here the Linux kernel cannot be
described as good. I won't say that I'm surprised because it's not the
first time you don't understand your claims.

>>
>> write something that does everything that
>> systemd is good at without his issues.
>>
>
> I already have written an init/service management system.

Yes, I know, it manages to display a useless black screen. You can be
proud of it.

> It is customized for my needs only.

Yes, your needs are only to be able to fail compiling your kernel.

> GNU/Linux makes this possible for me and for everyone else,

Yes, everyone else can have more expectation from his or her computer
than you. I have nothing against your limited needs/capabilities. The
issue with you is you consider others needs as invalids.

> but systemd threatens to destroy the possibility for everyone.

Nope. You just don't know how systemd works.

--
Si vous avez du temps à perdre :
https://scarpet42.gitlab.io

Le 19-05-2024, Farley Flud <ff@linux.rocks> a écrit :
>
> The systemd movement can be stopped

Yes.

> but it would take strong protests and

No, it would take a good alternative. For now, there is nothing serious.
Every alternative is either bad (like SysVInit) or too limited (they
just avoid the issues instead of correcting them).

So, like you pretend to be the best programmer in the world, you can
improve your init script until it does everything that systemd does or
you provide a better way to handle the things systemd shouldn't do.

> most users are totally subservient distro lackeys.

Nope. Either they don't care or they like it. The others, like you, are
fighting a war they don't understand. You only can claim it's bloated
without being able to show a better way.

--
Si vous avez du temps à perdre :
https://scarpet42.gitlab.io

On 19 May 2024 16:18:01 GMT, Stéphane CARPENTIER wrote:

>
> Not only and some softwares needs services. For the not only parts, I
> want to be able to [un]plug easily a USB key/drive, a mouse, a printer, the
> WiFi, well anything I didn't plugged before starting my computer.
>

Ha, ha, ha, ha, ha, ha!

I just plug in a USB drive and then invoke my own appropriate
script:

mount-usb-ext3

mount-usb-vfat

mount-scanner

Etc., etc.

(Printer? Only an idiot uses a printer.)

There is NO need for ridiculous services.

But YOU could not do this. YOU are totally dependent on systemd
services.

>
> for the softwares, when they run in docker, the service need to be
> started. A lot of things like the logs are very well managed thanks to a
> service.
>

Only a brain-dead, ignoramus retard would use "docker."

The case is closed.

>
> For example I remember when a few weeks/months ago you had issues with
> your ntpd service. You are the only one I know who were able to break
> it. That tells a lot.
>

You are hallucinating.

I don't have or need a ntpd service.

After boot I set my clock using the commands:

openrdate -4 -v -n pool.ntp.org

hwclock --systohc --utc --noadjfile

A total ignoramus and distro lackey like you cannot understand the
methods of a Linux master like me.

>>
>> Anything that requires 3 million lines of code to do
>> its job cannot be described as "good."
>
> So you are publicly telling everyone here the Linux kernel cannot be
> described as good. I won't say that I'm surprised because it's not the
> first time you don't understand your claims.
>

My kernel does not have anywhere near 3 million locs.

My kernel is heavily customized and therefore greatly reduced in size
form stock distro kernels.

>
> Yes, I know, it manages to display a useless black screen. You can be
> proud of it.
>

It's the KISS principle.

KISS = Keep it simple stupid.

>
> Nope. You just don't know how systemd works.
>

You don't know how Linux works.

Le 19-05-2024, Farley Flud <fflud@gnu.rocks> a écrit :
> On 19 May 2024 16:18:01 GMT, Stéphane CARPENTIER wrote:
>
>> for the softwares, when they run in docker, the service need to be
>> started. A lot of things like the logs are very well managed thanks to a
>> service.
>>
>
> Only a brain-dead, ignoramus retard would use "docker."

It's the future. It's to difficult for you to understand its use. You
are obsolete.

>> For example I remember when a few weeks/months ago you had issues with
>> your ntpd service. You are the only one I know who were able to break
>> it. That tells a lot.
>>
>
> You are hallucinating.
>
> I don't have or need a ntpd service.
>
> After boot I set my clock using the commands:
>
> openrdate -4 -v -n pool.ntp.org
>
> hwclock --systohc --utc --noadjfile
>
>
> A total ignoramus and distro lackey like you cannot understand the
> methods of a Linux master like me.

You explained very well how it didn't work. Looking for an explanation
in the system when the only needed explanation was you.

>>> Anything that requires 3 million lines of code to do
>>> its job cannot be described as "good."
>>
>> So you are publicly telling everyone here the Linux kernel cannot be
>> described as good. I won't say that I'm surprised because it's not the
>> first time you don't understand your claims.
>>
>
> My kernel does not have anywhere near 3 million locs.

You don't know how to count.

> My kernel is heavily customized and therefore greatly reduced in size
> form stock distro kernels.

I don't believe you. You are unable to customise anything. It wouldn't
work on your computer. Well, agreed, it doesn't work on your computer,
so maybe it's true. Make it work and it will have more lines of code
than that.

>> Yes, I know, it manages to display a useless black screen. You can be
>> proud of it.
>>
>
> It's the KISS principle.
>
> KISS = Keep it simple stupid.

It must be kept simple, but not to simple to start being useless.

>> Nope. You just don't know how systemd works.
>
> You don't know how Linux works.

Better than you who only know how Linux doesn't work.

--
Si vous avez du temps à perdre :
https://scarpet42.gitlab.io

On Sun, 19 May 2024 09:03:06 -0000 (UTC), Sebastian wrote:

> Even Devuan, whose reason for existence is to be a
> Systemd-free Debian, had to use a fork of systemd-logind rather than
> just handling sessions the pre-Systemd way, because so many things have
> been made to depend on it.

Surely they could come up with their own system for handling login
sessions. Just start with the code from before systemd came along.

On Tue, 25 Jun 2024 05:54:34 -0000 (UTC), Lawrence D'Oliveiro wrote:

> On Sun, 19 May 2024 09:03:06 -0000 (UTC), Sebastian wrote:
>
>> Even Devuan, whose reason for existence is to be a Systemd-free Debian,
>> had to use a fork of systemd-logind rather than just handling sessions
>> the pre-Systemd way, because so many things have been made to depend on
>> it.
>
> Surely they could come up with their own system for handling login
> sessions. Just start with the code from before systemd came along.

The problem is that Devuan is still based on Debian, so they have to
constantly take new Systemd-based packages from Debian and make them
work without Systemd. They will probably be overwhelmed at some point,
and then the entire distro will collapse. The same will probably happen
to other non-Systemd Linux distros. Those of us who still refuse to
run Systemd will have no choice but to install a BSD operating system.
I'll probably do that sooner rather than later.

On Tue, 25 Jun 2024 06:33:14 -0000 (UTC), Sebastian Wells wrote:

> On Tue, 25 Jun 2024 05:54:34 -0000 (UTC), Lawrence D'Oliveiro wrote:
>
>> On Sun, 19 May 2024 09:03:06 -0000 (UTC), Sebastian wrote:
>>
>>> Even Devuan, whose reason for existence is to be a Systemd-free
>>> Debian, had to use a fork of systemd-logind rather than just handling
>>> sessions the pre-Systemd way, because so many things have been made to
>>> depend on it.
>>
>> Surely they could come up with their own system for handling login
>> sessions. Just start with the code from before systemd came along.
>
> The problem is that Devuan is still based on Debian, so they have to
> constantly take new Systemd-based packages from Debian and make them
> work without Systemd. They will probably be overwhelmed at some point,
> and then the entire distro will collapse.

I wonder why that would happen? I thought there was a huge crowd of people
who hated systemd, who would be happy to support such a distro. Isn’t that
the case?

> Those of us who still refuse to run Systemd will have no choice but to
> install a BSD operating system.

Fun fact: some in the BSD world are working on their own systemd-
lookalike. It’s called “InitWare”.

On Tue, 25 Jun 2024 05:54:34 -0000 (UTC)
Lawrence D'Oliveiro <ldo@nz.invalid> wrote:
> Surely they could come up with their own system for handling login
> sessions. Just start with the code from before systemd came along.

There is one, it's called ConsoleKit2. All the systemd-free distros
ditched it because it didn't cover the "seat management" needed to
support Wayland sessions. These days we can use seatd for that. That
way ConsoleKit2 does the session tracking and seatd manages seats, no
need for an octopus like elogind covering both because both of those
components do their jobs very well.

It pains me how distros like Slackware or Devuan conceded to using
logind. Devuan at least still has the package and it's up to date, but
nothing is built to support it.

PCLinuxOS is the only distro I know still using Ck2 and I recently
ported seatd over to finally get working Wayland sessions on there.

On Wed, 26 Jun 2024 20:08:41 -0500, Hunter wrote:

> It pains me how distros like Slackware or Devuan conceded to using
> logind. Devuan at least still has the package and it's up to date, but
> nothing is built to support it.

Why not get some like-minded systemd-haters together to support it? There
must be no shortage of those, surely.

On Thu, 27 Jun 2024 07:44:46 -0000 (UTC)
Lawrence D'Oliveiro <ldo@nz.invalid> wrote:
>
> Why not get some like-minded systemd-haters together to support it?
> There must be no shortage of those, surely.

There are two camps. There are those who think removing systemd as the
init is enough, and others who understand there is more to systemd than
just the init. Most fall in the first category because it seems
daunting to try and replace anything else. Realistically all that
Slackware (just as an example) would need is to bring back ConsoleKit,
introduce seatd, and rebuild polkit and Network Manager. All parts of
systemd can be easily replaced except for udev, and that's unfortunate,
but at least udev was its own things for years beforehand and I guess
we could try something like mdevd.