Hi there

[Ian Jackson]

KatolaZ writes ("Hi there"):
> 1) as you know, at Devuan we are willing to help with/take over
> maintenance of sysvinit in Debian. I am currently reviewing the state
> of things (including bug reports and new development upstream), and
> will post here a report and a plan ASAP. I guess most of the
> complaints are indeed not on sysvinit but on the scripts provided by
> maintainers of other packages. We have had such issues with zfs (which
> lacks proper sysvinit scripts), and the matter has been escalated
> upstream (Debian) and will be apparently fixed in zfs 0.8*. Other
> packages have had similar issues recently.

Thanks.

> 2) elogind is already working in Devuan ascii, and all the desktop
> things seem to be fine with that. Please notice that the ceres
> (unstable) version is lagging behind, mainly because most of the
> development and bug fixing happened in ascii, to enable us to release
> it sooner. But we are coming back to the proper
> experimental/unstable/testing cycle now.

Should someone try the ascii version of elogind in buster maybe ?  Or
should we want for ceres to be updated.

> You might have noticed that the trick is to let elogind Provides:
> libpam-systemd, and a few more hacks around that. Unfortunately,
> supporting both elogind and consolekit proved a bit tricky, and
> required to provide two sets of polkit-1-* packages (one for the
> elogind backend and one for the consolekit one). But it should be
> possible to simplify stuff there.

I'm afraid I don't know anything about this.

> 3) You might also have noticed that ascii has eudev working just fine.

I understand why Devuan would want eudev rather than udev, but is
there actually likely to be a practical problem with systemd's udev,
and sysvinit, in buster ?  It works fine for me in stretch and I'm not
aware of reported problems.

> 4) We are working to include as many init/supervisors choices at
> install time as possible. Currently ASCII has also openrc, but the
> idea is to add more options in Beowulf.

That is nice.  I think on the Debian side we do have work to do on the
installer (and maybe we need to provide alternative an installer
build) but we need to get the installed system working right first.

Regards,
Ian.