[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [draft] Debian/Hurd porters position in the default init system debate



Hi,

Quoting Steven Chamberlain (2014-01-30 16:14:44)
> Since you didn't mention Upstart in this draft, may I ask your point of
> view on it?  Is there any prospect/interest in using it on GNU/Hurd?

I deliberately did not express any preference or mention neither
upstart nor systemd.  I'll update the draft to make that explicit:

~~ snip ~~~

We, the Debian/Hurd porters and developers, like to state our position
in the current debate about the default init system for Debian.

0. We have no preference for either candidate for the default init
system used by Debian.

1. Up to this day, Debian/Hurd has never used the current default init
system (sysvinit) but has relied on its own init and rc system.  We
are prepared to use a non-default init system in the future.

2. We are currently switching to sysvinit for Debian/Hurd.  All the
Hurd patches are in place, patches for the sysvinit package are
awaiting inclusion.

3. We ask that the current sysvinit-compatible init scripts are left
in place, so that we can use sysvinit in the future.

4. We acknowledge that there is a maintenance cost involved with
keeping the current init scripts.  We will help maintain them as part
of our porting effort.

5. We look forward to using OpenRC as incremental improvement.  OpenRC
complements sysvinit by replacing its rc component.  Work has started
to port OpenRC to Debian/Hurd and is progressing nicely.

~~ snip ~~~

Supporting an init system that is not portable (and I don't believe
any init system is) and is written for another platform is a lot of
work.  We have done quite some work over the years to make the Hurd
compatible enough to Linux (and sometimes we even have to cheat and
lie a little to that end) to run various Linux software, including
sysvinit.

Eventually the Hurd will be compatible enough to run upstart or
systemd if enough effort is put into that.  I have no idea if this
will ever happen.

> What if it becomes the default, or the second-most popular init system
> in Debian?  Or if it is used by GNU/kFreeBSD for jessie?

See point 1.

> The main benefit I see could be already-written job files, coming from
> Debian or Ubuntu, particularly if the package maintainer stops
> maintaining the sysvinit scripts.

What is the benefit of already-written upstart job files over
already-written sysvinit-compatible init scripts?  Or over
already-written systemd service files for that matter?

Cheers,
Justus


Reply to: