seeking advise: runit-init and 'protected: yes'

[Matthew Vernon]

Hi,

Let me be clear: the opinions that follow are my personal ones, not to 
be taken as representative of the TC (not least, I wasn't on the 
committee when the relevant discussion happened!).

On 16/03/2025 23:19, Lorenzo wrote:

> As sysvinit-core and systemd-sysv already did, I would like to add
> protected:yes to runit-init.
> By looking at #940965 and #872587 this looks correct.
> However runit was already involved in an escalation with a related issue
> that ended with a TC ruling (see #923450) and Michael says that init
> metapackage becomes moot and they are going to drop it(?)
> I'm not sure how to apply the TC ruling in the new context where
> each init-package gets 'protected: yes'. [1]

I think if you think it's appropriate to add protected: yes to 
runit-init, then you can do so. I don't think the TC bug you refer to is 
really relevant; it was about how much gatekeeping it was appropriate 
for the init metapackage maintainers to do. I think if you wanted to try 
and get runit-init into the dependencies of the init metapackage, it'd 
be a different conversation (and particularly, whether the bugs in runit 
that were identified in #838480 got fixed).

While the policy discussion on protected: yes hasn't concluded in 
#872587 it still seems to me that runit-init is within the 
likely-intended scope of protected: yes - if someone has installed it as 
their init system, they want it to be hard to accidentally remove it again.

Regards,

Matthew