Le 22/06/2024 à 10:44, Tomasz Torcz a écrit :
> Dnia Sat, Jun 22, 2024 at 10:20:44AM +0200, Martin Steigerwald napisał(a):
>> Nah, I better stop giving Systemd developers more ideas to put into the
>> Systemd system building blocks as I remember Lennart used to describe
>> Systemd with. That is also an interesting development: According to my
>> memory Lennart advertised Systemd as a init system replacement initially,
>> yet later wrote something like "no no, it was never mean just as an init
>> system, but as system building blocks". Nowadays Systemd fans usually
>> write that Systemd was never meant to be just an init system. As far as I
>> remember Lennart advertised it as just that initially. Not taking the time
>> to dig that out at the moment though. So take with a grain of salt. Does
>> anyone else remember it that way and may have suitable links proving it?
> Original announcement of systemd from 2010 is at https://0pointer.de/blog/projects/systemd.html
> At quick glance, it mentions process management capabilities only.
> This doesn't preclude extending the scope later, of course.
>
Interesting. Actually despite the fisrt word, all what is
described does not need any of the pecularities of PID1.
First it appears as a kind of INETD server; then they add the use
of AUTOFS, a feature created by RH, which allows the use of filesystems
before they are mounted.
I must say that, even on kernel.org, the description of AUTOFS is
very badly written, putting configuration options before explanations of
what is done. There is no real effort to make the explanations simple
and clear, which is typical of the team behind.
Anyway it was all about a startup system which includes PID1 for
no given reason, and it was not yet about a new OS on top of Linux.
-- Didier
::
Re: [DNG] Yeah,
what can even poss…
