Debian Bug report logs - #756023
init: Move "Essential: yes" from init to init-system-helpers

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Dimitri John Ledkov <xnox@ubuntu.com>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: init: Please drop Essential:yes from init metapackage

Date: Fri, 25 Jul 2014 15:19:37 +0100

[Message part 1 (text/plain, inline)]

Package: init
Version: 1.20
Severity: serious
Tags: patch

Dear Maintainer,

"init" metapackage is marked essential, even though none of the inits
are currently essential. There are environments in which no init is
needed, e.g. chroots and minimal/exec-only containers (like
docker). Also one shouldn't be making 3 inits transitively essential,
when they currently are not.

Patch attached.

[0001-Drop-Essential-yes-from-init-metapackage.-None-of-th.patch (text/x-diff, inline)]

From 99cde7b65ed839df2250f9278847ca20c64e6fd6 Mon Sep 17 00:00:00 2001
From: Dimitri John Ledkov <xnox@ubuntu.com>
Date: Fri, 25 Jul 2014 15:15:33 +0100
Subject: [PATCH] Drop "Essential: yes" from "init" metapackage. None of the
 init's are currently essential, and there is no use for init in many
 environments, e.g. chroots / exec-only containers (docker).

---
 debian/changelog | 8 ++++++++
 debian/control   | 1 -
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/debian/changelog b/debian/changelog
index e01de73..8d68f82 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,11 @@
+init-system-helpers (1.21) UNRELEASED; urgency=medium
+
+  * Drop "Essential: yes" from "init" metapackage. None of the init's are
+    currently essential, and there is no use for init in many
+    environments, e.g. chroots / exec-only containers (docker).
+
+ -- Dimitri John Ledkov <xnox@ubuntu.com>  Fri, 25 Jul 2014 15:14:34 +0100
+
 init-system-helpers (1.20) unstable; urgency=medium
 
   [ Michael Stapelberg ]
diff --git a/debian/control b/debian/control
index c9fe0df..ef8beec 100644
--- a/debian/control
+++ b/debian/control
@@ -41,7 +41,6 @@ Description: debhelper add-on to handle systemd unit files
 Package: init
 Section: metapackages
 Architecture: any
-Essential: yes
 Priority: required
 Pre-Depends: sysvinit-core | systemd-sysv [linux-any] | upstart [linux-any]
 Description: System-V-like init utilities - metapackage
-- 
2.0.1

[Message part 3 (text/plain, inline)]

Regards,

Dimitri.

Message #10 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Ansgar Burchardt <ansgar@debian.org>

To: Dimitri John Ledkov <xnox@ubuntu.com>, 756023@bugs.debian.org

Subject: Re: Bug#756023: init: Please drop Essential:yes from init metapackage

Date: Fri, 25 Jul 2014 16:28:42 +0200

Control: severity -1 wishlist

On 07/25/2014 16:19, Dimitri John Ledkov wrote:
> "init" metapackage is marked essential, even though none of the inits
> are currently essential. There are environments in which no init is
> needed, e.g. chroots and minimal/exec-only containers (like
> docker). Also one shouldn't be making 3 inits transitively essential,
> when they currently are not.

They are already transitively essential:

   Package: sysvinit
   Version: 2.88dsf-53.2
   Essential: yes
   [...]
   Pre-Depends: sysvinit-core | upstart | systemd-sysv

init is intended to replace the sysvinit package; it doesn't change
anything about init systems being essential or not.

Ansgar

Message #17 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Michael Biebl <biebl@debian.org>

To: Dimitri John Ledkov <xnox@ubuntu.com>, 756023@bugs.debian.org

Cc: Debian Bug Control <control@bugs.debian.org>

Subject: Re: Bug#756023: init: Please drop Essential:yes from init metapackage

Date: Fri, 25 Jul 2014 16:32:20 +0200

[Message part 1 (text/plain, inline)]

severity 756023 wishlist
thanks

Hi,

Am 25.07.2014 16:19, schrieb Dimitri John Ledkov:
> Package: init
> Version: 1.20
> Severity: serious
> Tags: patch
> 
> Dear Maintainer,
> 
> "init" metapackage is marked essential, even though none of the inits
> are currently essential. There are environments in which no init is
> needed, e.g. chroots and minimal/exec-only containers (like
> docker). Also one shouldn't be making 3 inits transitively essential,
> when they currently are not.

Well, see the transition plan as outlined and discussed on debian-devel
[1]. The new "init" metapackage takes over the role of the sysvinit
package, which (currently) is Essential: yes and Priority: required.
The point here is to ensure that an init system is installed at all
times. Dropping the Essential flag would counter that.

The question if "init" needs to be Essential came up as well.
The current sysvinit package in wheezy is Essential so packages do rely
on tools like /sbin/runlevel to be existing and working. Dropping the
Essential flag would break that assumption and as Tollef said would
require an archive wide cleanup before we could drop this flag.

I'm marking this as a whishlist bug as I see no justitification for it
being serious.

Regards,
Michael


[1] https://lists.debian.org/debian-devel/2014/07/msg00611.html
[2] https://lists.debian.org/debian-devel/2014/07/msg00631.html
-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

[signature.asc (application/pgp-signature, attachment)]

Message #22 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Dimitri John Ledkov <xnox@ubuntu.com>

To: Ansgar Burchardt <ansgar@debian.org>

Cc: 756023@bugs.debian.org

Subject: Re: Bug#756023: init: Please drop Essential:yes from init metapackage

Date: Fri, 25 Jul 2014 15:48:09 +0100

On 25 July 2014 15:28, Ansgar Burchardt <ansgar@debian.org> wrote:
> Control: severity -1 wishlist
>
> On 07/25/2014 16:19, Dimitri John Ledkov wrote:
>> "init" metapackage is marked essential, even though none of the inits
>> are currently essential. There are environments in which no init is
>> needed, e.g. chroots and minimal/exec-only containers (like
>> docker). Also one shouldn't be making 3 inits transitively essential,
>> when they currently are not.
>
> They are already transitively essential:
>
>    Package: sysvinit
>    Version: 2.88dsf-53.2
>    Essential: yes
>    [...]
>    Pre-Depends: sysvinit-core | upstart | systemd-sysv
>
> init is intended to replace the sysvinit package; it doesn't change
> anything about init systems being essential or not.
>

In that case, apart from a new name and new source-package, why do we
need two identical "metapackages"?
This seems utterly redundant and more complicated then simply using
sysvinit package for the transition. Ok, whatever.

-- 
Regards,

Dimitri.

Message #27 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Michael Biebl <biebl@debian.org>

To: Dimitri John Ledkov <xnox@ubuntu.com>, 756023@bugs.debian.org

Subject: Re: Bug#756023: init: Please drop Essential:yes from init metapackage

Date: Fri, 25 Jul 2014 16:54:54 +0200

[Message part 1 (text/plain, inline)]

Am 25.07.2014 16:48, schrieb Dimitri John Ledkov:

> In that case, apart from a new name and new source-package, why do we
> need two identical "metapackages"?
> This seems utterly redundant and more complicated then simply using
> sysvinit package for the transition. Ok, whatever.

Well, read the transition plan I referenced in my other reply.

The point is, that we want different behaviour depending on whether you
upgrade or do a fresh installation.

The sysvinit package in jessie won't be Essential: yes / Priority:
required anymore (upload for that is pending).

It will become an (optional) transitional package with the only purpose
being to ensure a smooth upgrade and providing a safety net.

Hope that clears things up.

Michael

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

[signature.asc (application/pgp-signature, attachment)]

Message #32 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Jan Braun <janbraun@gmx.net>

To: Debian Bug Tracking System <756023@bugs.debian.org>

Subject: Re: init: Please drop Essential:yes from init metapackage

Date: Wed, 27 Aug 2014 18:21:41 +0200

[Message part 1 (text/plain, inline)]

Hi,
since init with essential:yes entered testing recently, I had a surprise
when apt-get on my STABLE system today suddenly wanted to install
systemd and dependencies[1][2].

Reason being that I had both stable and testing lines in my sources.list
and apt-pinning set up to prefer stable (and no testing packages
actually installed). apt-get considers an essential:yes package in any
suite to be essential for the complete system. See #216768, which
explains to my satisfaction how that is the most correct, if surprising,
behaviour.[3]

I think that would have been a good reason to keep sysvinit the
essential:yes package, even if it no longer contained System V init.
Of course, that may be irrelevant by now.


I don't necessarily have a point, just wanted to document that
experience for consideration by those manipulating essential:yes
packages.

regards,
    Jan

[1]
$ sudo apt-get dist-upgrade
Reading package lists... Done
Building dependency tree
Reading state information... Done
Calculating upgrade... Done
The following NEW packages will be installed:
  libaudit0 libsystemd-daemon0 libsystemd-id128-0 libsystemd-journal0 systemd
0 upgraded, 5 newly installed, 0 to remove and 0 not upgraded.
Need to get 0 B/1547 kB of archives.
After this operation, 3943 kB of additional disk space will be used.
Do you want to continue [Y/n]? ^C

Even now, I don't understand this completely, as neither init nor
systemd-sysv were going to be installed...

[2]
My own debugging was not helped by the coincidence of DSA-3012-1 updating
eglibc at the same time. Thanks to #debian-systemd for helping me
figuring out what was going on.

[3]
It's worth noting that aptitude seems to disagree and would not install
systemd.

[signature.asc (application/pgp-signature, inline)]

Message #37 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Ansgar Burchardt <ansgar@debian.org>

To: 756023@bugs.debian.org

Subject: init: Move "Essential: yes" from init to init-system-helpers

Date: Thu, 05 May 2016 15:05:37 +0200

Control: retitle -1 init: Move "Essential: yes" from init to init-system-helpers
Control: tag -1 - patch
Control: forcemerge -1 823501

I would like "init" to be optional in Debian 9 for chroot environments
and some uses of containers.  A first step seems to be making "init" no
longer essential; it would nice nice if the priority could later be
downgraded as well (to "important") so that a minimal debootstrap will
not install it.

As maintainer scripts call `invoke-rc.d` and `update-rc.d`
unconditionally, the package providing these scripts should be made
essential (i.e. add "Essential: yes" to init-system-helpers).

So, suggestion to proceed:

 - Now:
   + "init" package: Remove "Essential: yes"
   + "init-system-helpers" package: Add "Essential: yes"
 - Later:
   + "init": Change priority from "required" to "important".

Do we miss anything to start with moving the "Essential" field?  I think
/bin/pidof was mentioned which is provided by sysvinit-utils; however
sysvinit-utils will continue to stay quasi-essential as util-linux has
  Depends: sysvinit-utils (>= 2.88dsf-59.1~)
and util-linux is an essential package.

And do we miss anything for the priority change after that (besides
confirming with d-boot@)?

Ansgar

Message #48 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Felipe Sateler <fsateler@debian.org>

To: Ansgar Burchardt <ansgar@debian.org>, 756023@bugs.debian.org

Subject: Re: Bug#756023: init: Move "Essential: yes" from init to init-system-helpers

Date: Thu, 5 May 2016 10:18:36 -0300

On 5 May 2016 at 10:05, Ansgar Burchardt <ansgar@debian.org> wrote:
> Control: retitle -1 init: Move "Essential: yes" from init to init-system-helpers
> Control: tag -1 - patch
> Control: forcemerge -1 823501
>
> I would like "init" to be optional in Debian 9 for chroot environments
> and some uses of containers.  A first step seems to be making "init" no
> longer essential; it would nice nice if the priority could later be
> downgraded as well (to "important") so that a minimal debootstrap will
> not install it.
>
> As maintainer scripts call `invoke-rc.d` and `update-rc.d`
> unconditionally, the package providing these scripts should be made
> essential (i.e. add "Essential: yes" to init-system-helpers).

Just thinking out loud, but maybe it would be reasonable to provide
dummy implementations for container/chroot uses. If the real
implementation lives in the Essential: yes package, that makes it more
difficult to swap them.

Of course, that indirection package need not be init.

>
> So, suggestion to proceed:
>
>  - Now:
>    + "init" package: Remove "Essential: yes"
>    + "init-system-helpers" package: Add "Essential: yes"
>  - Later:
>    + "init": Change priority from "required" to "important".

Later, as in post-stretch?


> Do we miss anything to start with moving the "Essential" field?  I think
> /bin/pidof was mentioned which is provided by sysvinit-utils; however
> sysvinit-utils will continue to stay quasi-essential as util-linux has
>   Depends: sysvinit-utils (>= 2.88dsf-59.1~)
> and util-linux is an essential package.

The util-linux maintainer is very much trying to get rid of that
dependency (I can't find an online reference, but discussed on IRC),
and it is not done only because apt currently does not deal
particularly well with versioned Breaks "loops" involving essential
packages.

-- 

Saludos,
Felipe Sateler

Message #53 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Felipe Sateler <fsateler@debian.org>

To: Ansgar Burchardt <ansgar@debian.org>, 756023@bugs.debian.org

Subject: Re: Bug#756023: init: Move "Essential: yes" from init to init-system-helpers

Date: Thu, 5 May 2016 10:32:34 -0300

On 5 May 2016 at 10:18, Felipe Sateler <fsateler@debian.org> wrote:
>> Do we miss anything to start with moving the "Essential" field?  I think
>> /bin/pidof was mentioned which is provided by sysvinit-utils; however
>> sysvinit-utils will continue to stay quasi-essential as util-linux has
>>   Depends: sysvinit-utils (>= 2.88dsf-59.1~)
>> and util-linux is an essential package.
>
> The util-linux maintainer is very much trying to get rid of that
> dependency (I can't find an online reference, but discussed on IRC),
> and it is not done only because apt currently does not deal
> particularly well with versioned Breaks "loops" involving essential
> packages.


And I found it: it was tried and reverted
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=786469

-- 

Saludos,
Felipe Sateler

Message #58 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Ansgar Burchardt <ansgar@debian.org>

To: 756023@bugs.debian.org

Subject: Re: Bug#756023: init: Move "Essential: yes" from init to init-system-helpers

Date: Thu, 05 May 2016 15:49:57 +0200

Felipe Sateler writes:
> On 5 May 2016 at 10:05, Ansgar Burchardt <ansgar@debian.org> wrote:
>> As maintainer scripts call `invoke-rc.d` and `update-rc.d`
>> unconditionally, the package providing these scripts should be made
>> essential (i.e. add "Essential: yes" to init-system-helpers).
>
> Just thinking out loud, but maybe it would be reasonable to provide
> dummy implementations for container/chroot uses. If the real
> implementation lives in the Essential: yes package, that makes it more
> difficult to swap them.
>
> Of course, that indirection package need not be init.

The indirection package is currently "init-system-helpers".  The
implementation provided there should work with all init systems (and
hopefully also with no init system).

>> So, suggestion to proceed:
>>
>>  - Now:
>>    + "init" package: Remove "Essential: yes"
>>    + "init-system-helpers" package: Add "Essential: yes"

Oh, and another idea: apt understand an "Important" field that makes it
warn when removing packages (the "I know what I am doing" prompt).  It
might make sense to add this to either "init" or the actual init
systems so people don't uninstall an init system or switch to another by
accident.

>>  - Later:
>>    + "init": Change priority from "required" to "important".
>
> Later, as in post-stretch?

No, after removing the "Essential" bit and confirming with d-boot@, but
hopefully before the next freeze.

>> Do we miss anything to start with moving the "Essential" field?  I think
>> /bin/pidof was mentioned which is provided by sysvinit-utils; however
>> sysvinit-utils will continue to stay quasi-essential as util-linux has
>>   Depends: sysvinit-utils (>= 2.88dsf-59.1~)
>> and util-linux is an essential package.
>
> The util-linux maintainer is very much trying to get rid of that
> dependency (I can't find an online reference, but discussed on IRC),
> and it is not done only because apt currently does not deal
> particularly well with versioned Breaks "loops" involving essential
> packages.

Hmm, then alternatives include:

  - Make "sysvinit-utils" essential (not so nice IMHO).
  - Have "init-system-helpers" depend on it. This makes sure it stays
    quasi-essential for now.

Ansgar

Message #63 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Andreas Henriksson <andreas@fatal.se>

To: Ansgar Burchardt <ansgar@debian.org>, Felipe Sateler <fsateler@debian.org>

Cc: 756023@bugs.debian.org

Subject: Re: Bug#756023: init: Move "Essential: yes" from init to init-system-helpers

Date: Fri, 6 May 2016 14:26:21 +0200

Hello all.

On Thu, May 05, 2016 at 03:49:57PM +0200, Ansgar Burchardt wrote:
> Felipe Sateler writes:
[...]
> > The util-linux maintainer is very much trying to get rid of that
> > dependency (I can't find an online reference, but discussed on IRC),
> > and it is not done only because apt currently does not deal
> > particularly well with versioned Breaks "loops" involving essential
> > packages.
> 
> Hmm, then alternatives include:
> 
>   - Make "sysvinit-utils" essential (not so nice IMHO).
>   - Have "init-system-helpers" depend on it. This makes sure it stays
>     quasi-essential for now.

As already mentioned on IRC but following up here for the record,
sysvinit-utils is actually already Essential: yes.

(Irrelevant, but I've now filed #823569 for switching util-linux
Depends on sysvinit-utils to a Breaks.)

Please also note that I'd like to see sysvinit-utils become
non-essential in a potential soon future. My investigations seems to
suggest that pidof is the only really widely used part of current
(>= stretch) sysvinit-utils, so I've filed #810018 where tracking of
potentially providing procps pidof as an essential package (eg.
procps-base). Other work needed to be able to demote sysvinit-utils to
non-essential are also mentioned in that bug report. (Hopefully
long-term we can even make pidof non-essential as mostly init scripts is
where it's widespead usage comes from, but I don't see those going away
any time soon.)

Regards,
Andreas Henriksson

Message #68 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Martin Pitt <mpitt@debian.org>

To: Ansgar Burchardt <ansgar@debian.org>, 756023@bugs.debian.org

Subject: Re: Bug#756023: init: Move "Essential: yes" from init to init-system-helpers

Date: Thu, 19 May 2016 18:11:10 +0200

Control: tag -1 pending

Hello all,

Ansgar Burchardt [2016-05-05 15:05 +0200]:
> I would like "init" to be optional in Debian 9 for chroot environments
> and some uses of containers.  A first step seems to be making "init" no
> longer essential; it would nice nice if the priority could later be
> downgraded as well (to "important") so that a minimal debootstrap will
> not install it.

We revisited that again last week on IRC, and confirmed that
invoke-rc.d and update-rc.d behave well enough in environments without
any init. This was the main thing I wanted to assert before we do
this, I'm not aware of any other reasons why this would not work any
more.

>  - Now:
>    + "init" package: Remove "Essential: yes"
>    + "init-system-helpers" package: Add "Essential: yes"
>  - Later:
>    + "init": Change priority from "required" to "important".

I committed both in

  http://anonscm.debian.org/cgit/collab-maint/init-system-helpers.git/commit/?id=da29f8242

The priority change won't actually take effect until the ftpmasters
change the overrides, so I don't think it hurts doing the change in
the source now.

> Do we miss anything to start with moving the "Essential" field?  I think
> /bin/pidof was mentioned which is provided by sysvinit-utils; however
> sysvinit-utils will continue to stay quasi-essential as util-linux has
>   Depends: sysvinit-utils (>= 2.88dsf-59.1~)

Not any more, but sysvinit-utils is essential by itself.

> And do we miss anything for the priority change after that (besides
> confirming with d-boot@)?

Would you mind starting the discussion on d-boot for that?

Thanks!

Martin
-- 
Martin Pitt                        | http://www.piware.de
Ubuntu Developer (www.ubuntu.com)  | Debian Developer  (www.debian.org)

Message #75 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Ansgar Burchardt <ansgar@debian.org>

To: 756023@bugs.debian.org

Subject: Re: Bug#756023: init: Move "Essential: yes" from init to init-system-helpers

Date: Thu, 19 May 2016 18:26:46 +0200

Hi,

On Thu, 2016-05-19 at 18:11 +0200, Martin Pitt wrote:
> > And do we miss anything for the priority change after that (besides
> > confirming with d-boot@)?
> Would you mind starting the discussion on d-boot for that?

I would like to wait a few more days (until the next d-i alpha
currently in preparation is out).  I already mentioned my plans on IRC
a few days ago (in #-release, but KiBi knows about it).

Ansgar

Message #80 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Michael Biebl <biebl@debian.org>

To: Martin Pitt <mpitt@debian.org>, 756023@bugs.debian.org, Ansgar Burchardt <ansgar@debian.org>

Subject: Re: Bug#756023: init: Move "Essential: yes" from init to init-system-helpers

Date: Thu, 19 May 2016 18:27:28 +0200

[Message part 1 (text/plain, inline)]

Am 19.05.2016 um 18:11 schrieb Martin Pitt:
> Control: tag -1 pending
> 
> Hello all,
> 
> Ansgar Burchardt [2016-05-05 15:05 +0200]:
>> I would like "init" to be optional in Debian 9 for chroot environments
>> and some uses of containers.  A first step seems to be making "init" no
>> longer essential; it would nice nice if the priority could later be
>> downgraded as well (to "important") so that a minimal debootstrap will
>> not install it.
> 
> We revisited that again last week on IRC, and confirmed that
> invoke-rc.d and update-rc.d behave well enough in environments without
> any init. This was the main thing I wanted to assert before we do
> this, I'm not aware of any other reasons why this would not work any
> more.
> 
>>  - Now:
>>    + "init" package: Remove "Essential: yes"
>>    + "init-system-helpers" package: Add "Essential: yes"
>>  - Later:
>>    + "init": Change priority from "required" to "important".
> 
> I committed both in
> 
>   http://anonscm.debian.org/cgit/collab-maint/init-system-helpers.git/commit/?id=da29f8242

Could we add the Important: yes field to "init"? Having that makes a lot
of sense to me.

Regards,
Michael


-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

[signature.asc (application/pgp-signature, attachment)]

Message #85 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Michael Biebl <biebl@debian.org>

To: Martin Pitt <mpitt@debian.org>, 756023@bugs.debian.org, Ansgar Burchardt <ansgar@debian.org>

Subject: Re: Bug#756023: init: Move "Essential: yes" from init to init-system-helpers

Date: Thu, 19 May 2016 18:29:13 +0200

[Message part 1 (text/plain, inline)]

Am 19.05.2016 um 18:27 schrieb Michael Biebl:
> Am 19.05.2016 um 18:11 schrieb Martin Pitt:

>>
>>>  - Now:
>>>    + "init" package: Remove "Essential: yes"
>>>    + "init-system-helpers" package: Add "Essential: yes"
>>>  - Later:
>>>    + "init": Change priority from "required" to "important".
>>
>> I committed both in
>>
>>   http://anonscm.debian.org/cgit/collab-maint/init-system-helpers.git/commit/?id=da29f8242
> 
> Could we add the Important: yes field to "init"? Having that makes a lot
> of sense to me.

Now that i-s-h is Essential: yes, we could drop the Depends: i-s-h from
the init meta package.

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

[signature.asc (application/pgp-signature, attachment)]

Message #90 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Ansgar Burchardt <ansgar@debian.org>

To: 756023@bugs.debian.org

Subject: Re: Bug#756023: init: Move "Essential: yes" from init to init-system-helpers

Date: Thu, 19 May 2016 18:29:35 +0200

On Thu, 2016-05-19 at 18:27 +0200, Michael Biebl wrote:
> Could we add the Important: yes field to "init"? Having that makes a
> lot of sense to me.

To init or to its dependencies (systemd-sysv, sysvinit-core)?  The
latter would not only warn when uninstalling all init systems, but also
when switching between them (perhaps not by intention).

Ansgar

Message #95 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Michael Biebl <biebl@debian.org>

To: Ansgar Burchardt <ansgar@debian.org>, 756023@bugs.debian.org

Subject: Re: Bug#756023: init: Move "Essential: yes" from init to init-system-helpers

Date: Thu, 19 May 2016 18:45:03 +0200

[Message part 1 (text/plain, inline)]

Am 19.05.2016 um 18:29 schrieb Ansgar Burchardt:
> On Thu, 2016-05-19 at 18:27 +0200, Michael Biebl wrote:
>> Could we add the Important: yes field to "init"? Having that makes a
>> lot of sense to me.
> 
> To init or to its dependencies (systemd-sysv, sysvinit-core)?  The
> latter would not only warn when uninstalling all init systems, but also
> when switching between them (perhaps not by intention).

To "init" I guess. I don't think we should pester users with that
message if they deliberately switch to an alternative.
The main objective of having Important:yes in init would be to avoid
"accidentally" uninstalling /sbin/init and rendering your system unbootable.

Michael


-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?

[signature.asc (application/pgp-signature, attachment)]

Message #100 received at 756023@bugs.debian.org (full text, mbox, reply):

From: Martin Pitt <mpitt@debian.org>

To: Michael Biebl <biebl@debian.org>

Cc: 756023@bugs.debian.org, Ansgar Burchardt <ansgar@debian.org>

Subject: Re: Bug#756023: init: Move "Essential: yes" from init to init-system-helpers

Date: Sun, 29 May 2016 11:34:27 +0200

[Message part 1 (text/plain, inline)]

Hello all,

Michael Biebl [2016-05-19 18:27 +0200]:
> Could we add the Important: yes field to "init"? Having that makes a lot
> of sense to me.

Indeed, I did that now in git. I tested in a schroot, and "apt-get
purge init" now indeed complains ("yes, do as I say"), but one can
still remove it with dpkg, and it should fall out of debootstrap.

This is really nice now -- out of essential, but prevents accidental
removal *if* you have it installed.

Martin
-- 
Martin Pitt                        | http://www.piware.de
Ubuntu Developer (www.ubuntu.com)  | Debian Developer  (www.debian.org)

[signature.asc (application/pgp-signature, inline)]

Message #105 received at 756023-close@bugs.debian.org (full text, mbox, reply):

From: Martin Pitt <mpitt@debian.org>

To: 756023-close@bugs.debian.org

Subject: Bug#756023: fixed in init-system-helpers 1.34

Date: Mon, 30 May 2016 16:24:11 +0000

Source: init-system-helpers
Source-Version: 1.34

We believe that the bug you reported is fixed in the latest version of
init-system-helpers, which is due to be installed in the Debian FTP archive.

A summary of the changes between this version and the previous one is
attached.

Thank you for reporting the bug, which will now be closed.  If you
have further comments please address them to 756023@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Martin Pitt <mpitt@debian.org> (supplier of updated init-system-helpers package)

(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmaster@ftp-master.debian.org)


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Format: 1.8
Date: Mon, 30 May 2016 15:52:48 +0200
Source: init-system-helpers
Binary: init-system-helpers dh-systemd init
Architecture: source all amd64
Version: 1.34
Distribution: unstable
Urgency: medium
Maintainer: Debian systemd Maintainers <pkg-systemd-maintainers@lists.alioth.debian.org>
Changed-By: Martin Pitt <mpitt@debian.org>
Description:
 dh-systemd - debhelper add-on to handle systemd unit files
 init       - init metapackage
 init-system-helpers - helper tools for all init systems
Closes: 756023 820359 825075
Changes:
 init-system-helpers (1.34) unstable; urgency=medium
 .
   [ Martin Pitt ]
   * Make "init" not essential any more and lower priority from "required" to
     "important", so that we can drop it from minimal chroots and docker
     containers at some point. Explicitly mark init-system-helpers as essential
     now (before it was only transitively essential due to  init depending on
     it). (Closes: #756023)
   * Mark init "Important: yes" instead, so that one does not accidentally
     uninstall it.
   * deb-systemd-invoke: Accept policy-rc.d exit code 0 as "allowed" in
     addition to code 104, like invoke-rc.d. (Closes: #825075)
 .
   [ Dan Nicholson ]
   * deb-systemd-helper, dh_systemd_start: Use split rather than
     Text::ParseWords::shellwords since the latter will strip valid \
     escapes from unit names. The values then need to have leading and
     trailing quotes stripped. (Closes: #820359)
   * dh_systemd_enable: Drop unused Text::ParseWords use.
Checksums-Sha1:
 8ec95355251f2a664fe634b3023e1f2f4d42ede2 1867 init-system-helpers_1.34.dsc
 4c2bad2b9da68c9c8499a7ec3a1b170713c51612 57696 init-system-helpers_1.34.tar.xz
 fc42b5c2d5eba3aaf87490398882afb7a5317f6e 21986 dh-systemd_1.34_all.deb
 ec30e7fe47b9ab171b4dad079013ac3befb85ce1 35772 init-system-helpers_1.34_all.deb
 5e1faa3e65a1500e33008ed16f982651c0f9ecfc 8162 init_1.34_amd64.deb
Checksums-Sha256:
 5db2f2c6bab48ba32d99f0ed1f3364576060c2be6ed65fc318a94f12341e582e 1867 init-system-helpers_1.34.dsc
 328bec32d5e95c2e81dfbb027bb3e2bfc847d8cf759f548010c1dd1934ece625 57696 init-system-helpers_1.34.tar.xz
 0980e6daf3a1faad5fdbe7cf87985c259043714387c7184dac9695d076b9852b 21986 dh-systemd_1.34_all.deb
 c36f7e594fa157afeef6be10ccb110f6b30696a820aca4d04330d82606f0d00a 35772 init-system-helpers_1.34_all.deb
 a00a40a2f8e08b3bfe42788cafaf6e32bbaa260cc6b6799e9c8f7eb9844388dc 8162 init_1.34_amd64.deb
Files:
 5ce18fb948eb3699234ab53f256820b4 1867 admin extra init-system-helpers_1.34.dsc
 608ac7c7194b13df0df8c3456cd62086 57696 admin extra init-system-helpers_1.34.tar.xz
 34f5f815857a2fb851bc8933f3e0c423 21986 admin extra dh-systemd_1.34_all.deb
 fa1b479dce3d9e5ac663dd48d095ceae 35772 admin required init-system-helpers_1.34_all.deb
 039da57b26ec2ae7e02444c3e7f1e5a3 8162 metapackages important init_1.34_amd64.deb

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
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=9I5p
-----END PGP SIGNATURE-----

Send a report that this bug log contains spam.