Bug#727708: OpenRC now works on GNU/Hurd! :)
Hi,
It's with great joy that I can announce here that OpenRC now supports
GNU/Hurd. I have just added a few patches which were worked out with
upstream (you can look at them, it's really trival FTBFS fixing...),
tested it, and I can happily say it works.
The only thing that bothers me a bit is that the ANSI output isn't so
pretty, but I guess this is fixable and a minor problem.
On the Thu, 16 Jan 2014 17:18:24 +0100, Josselin Mouette
<joss@debian.org> wrote:
> This assumes that OpenRC can be fixed to have parallel boot, otherwise
> this is a big regression from the current insserv setup.
This is just plain wrong, OpenRC perfectly supports parallel booting,
it's just that the output on the screen is very ugly for the moment
(that as well can be fixed, I suppose).
Also, I'd like to point out to everyone that the OpenRC runscripts are
stored in /etc/init.d. This means that if someone wants to support
OpenRC and use a runscript instead of a traditional init.d shell script,
that someone will also have to support whatever we will choose as
default. Let me explain to make sure everyone gets it...
Let's say you rewrite /etc/init.d/foo, and transform it from a init.d
traditional sysv-rc script to an OpenRC runscript in your package. If
the init system is systemd, then systemd will *not* understand the
OpenRC runscript. This means that you will also have to write a systemd
unit file, if you want to write /etc/init.d/foo as an OpenRC runscript.
The same would of course apply to Upstart.
Though I think that writing a systemd unit file, plus an OpenRC
runscript, is still more easy and strait forward than writing a single
init.d sysv-rc shell script.
So, if we are to switch to systemd as default (same would apply if we
choose Upstart), IMO the policy should be that package maintainers have
2 choices:
1/ Write a standard LSB-header SysV-rc init script, which will of course
work with any init system (sysv-rc, OpenRC, systemd, Upstart, file-rc...)
2/ If the /etc/init.d/foo is an OpenRC runscript (which we should, IMO,
push for since traditional scripts have some many problems which I can't
even lest in this mail, and we all agree about that), then the package
maintainer *must* provide support for systemd (or upstart, if we choose
that as default).
IMO, the above would be the best way forward for Debian if we want to
continue to support our ports.
Cheers,
Thomas Goirand (zigo)
Reply to: