[libre-riscv-dev] Tedious systemd bunfight.

Veera vklr at vkten.in
Sat Feb 15 17:29:17 GMT 2020


On Sat, Feb 15, 2020 at 10:41:02AM -0500, Hendrik Boom wrote:
> On Sat, Feb 15, 2020 at 03:08:31PM +0000, Luke Kenneth Casson Leighton wrote:
> > On Sat, Feb 15, 2020 at 2:41 PM Hendrik Boom <hendrik at topoi.pooq.com> wrote:
> > 
> 
> Most init systems are quite compatible with each other in this respect.
> Systemd isn't.
> 
> I would have preferred if systemd had remained just an init system and hadn't
> invasively caused the schism we see today.
> 

Perhaps, by requirement and design systemd could not have remained just a
simple init system. Redhat wanted something coherent, unifying thing for
their RHEL, etc. Simple init systems people argue about the inherent design
flaw of systemd that the process 1 should not be so big and do so many things.
They say a single bug or vulnerability will collapse the whole system.

For above reasons I used runit first and then s6 as the init system.

Debian chose systemd as it was left with no other easy option.
For example most other softwares switched to using systemd and it's
features as a hard requirement. For example udev was integrated into systemd
and then after some versions became a inseparable part from it. So eudev was
forked by Gentoo developers.

Regards,
Veera



More information about the libre-riscv-dev mailing list