Re: Re-Proposal - preserve freedom of choice of init systems
On 16 October 2014 23:07, Paul Tagliamonte <paultag@debian.org> wrote:
>> * if your software only works if started by this one init system -
>> that is a problem.
>
> I don't quite understand this - what if you depend on something that's
> only provided / supported on one init system? Take for example the case
> of logind before we had the shim?
According to my reading of the proposal - either logind gets an RC bug
for not being able to work with other init systems (here logind is
considered a separate piece of software) or apps who have a hard
dependency on logind (without fallbacks for cases when it does not
exist or is not running) get a RC bug for that (here logind is
considered to be an integral part of systemd).
>> The requirement is that software should be able to work regardless of
>> how it is started - by systemd, by sysvinit, by other init system or
>> by a plain shell script called from the "init=" kernel parameter. If
>> there are any dependant services, those should be also able to be
>> simply startable by anything.
>
> So we can't rely on a new library until that library is supported on all
> init systems? (e.g. logind before shim, etc)
I'd see it as - new library can't get into Debian [stable release]
until is capable of working with installations being managed by any
init system or even no init system at all (like inside a Docker
container ;)).
>> All software in previous Debian releases
>> satisfied this requirement, so there wasn't even any need to consider
>> adding such requirement to the policy.
>
>
> Getting tired of these threads :(
True that. But at least this is about an actual point. (IMHO)
--
Best regards,
Aigars Mahinovs mailto:aigarius@debian.org
#--------------------------------------------------------------#
| .''`. Debian GNU/Linux (http://www.debian.org) |
| : :' : Latvian Open Source Assoc. (http://www.laka.lv) |
| `. `' Linux Administration and Free Software Consulting |
| `- (http://www.aiteki.com) |
#--------------------------------------------------------------#
Reply to: