On 28/05/26 8:53 pm, Matthias Geiger wrote:
> On Thu, 28 May 2026 17:23, Ansgar 🙀 <[email protected]> wrote:
>> Hi,
>>
>> On Thu, 2026-05-28 at 16:56 +0200, Matthias Geiger wrote:
>>> On Thu, 28 May 2026 16:47, Ansgar 🙀 <[email protected]> wrote:
>>>> It seems wrong to use /etc/init.d for a different purpose than it is
>>>> intended for.
>>>>
>>> As mention in my other bug, openrc does support /etc/init.d as main
>>> location. I assume this needs policy changed; nevertheless this a
>>> bug.
>>
>> No, this sounds like it really needs openrc changed.
>>
>> Or support for LSB init scripts removed completely and replaced by
>> openrc. That would work too.
>>
> Sorry, let clear this up a bit: I have been working with the openrc
> maintainer to create a openrc-native package that successfully can boot,
> shutdown, and restart a Debian system while *without* using any sysvinit
> scripts or dependencies. While this change is not in the archive yet,
> this already has been tested in qemu to be working.
>
> I can block those bug in the meantime by the ccorresponding openrc one.
Lintian can not go against policy. The policy currently says things more
or less explicitly pertaining to sysvinit, and is unclear on what openrc
supporting init.d should (or should not) mean for the packages using the
same.
If there are disagreements on how this should be handled, I would not want
lintian to recommend things otherwise.
Please file a bug for policy and add that as a blocking bug for this bug.
Once the policy is amended, or clarified by the policy maintainers, feel
free to give me a ping.
Please do so for #1138165 #1138169 and #1138167 that you filed earlier today
as well.
Thanks for working on openrc!
Best,
Nilesh
