On Fri, 19 Apr 2019 12:20:52 +0200
aitor_czr <aitor_czr@???> wrote:

> So, i'll give here
> my opinion: any risky decision (as well as a replacement to sysvinit)
> should be
> taken in a devuan derivative and not in the devuan operanting system
> per se,which is a base to build upon.

Sounds good to me.

As per sysvinit replacement, as the resident sysvinit detractor and
resident runit/s6 fan, I see no reason to hurry the replacement of
sysvinit, at least as an init system.

To my knowledge, there's nothing wrong with the "stage 1" part of
sysvinit: The part that brings up the OS to the point where daemons can
be run. Nor do I have a problem with the PID1 part of sysvinit. My beef
with sysvinit is only the part that handles daemons.

That part of sysvinit can be replaced by runit or s6, *by the
computer's user*. All we need is to make sure that the Devuan packages
for runit *process supervision* and s6 *process supervision* work
correctly when installed. Then the user can install runit or s6 and
slowly transfer daemons from sysvinit to runit or s6, at his or her
leisure.

I haven't fulfilled my promise to 1) make an s6 supervision package and
2) Hand over a collection of curated run files for each of s6 and
runit. I'm sorry about that --- I've had a lot of time management
problems the past two years. I'll try to do at least a partial
collection of #2 in the next half-year. If anybody who is faster or has
more time than I does some of this stuff before me, especially #1, I
won't be insulted: I'll be relieved.

In summary, there's absolutely no reason to be in any rush to replace
sysvinit: For the foreseeable future, the (incomplete) process
supervision of sysvinit can be informally replaced by runit or s6, very
easily, by a motivated user. I can help such a user.


SteveT