Updating inits

[anticapitalistaForum Admin]

This is for antiX-26 trixie users, not testing or sid.

Should init(s) get updates?
There are updates to sysVinit and s6.
I successfully run both updated packages on my frugal trixie installs (both arches)
and the latest s6 on my antiX-23 installed system.

Philosophers have interpreted the world in many ways; the point is to change it.

antiX with runit - leaner and meaner.

[andfreeMember]

anticapitalista wrote:

Should init(s) get updates?

Why not? Is there any specific reason?
From a more general point of view, I would like to have

andfree wrote:

the option of not upgrading the hidden inits as long as they remain hidden

but I assume that it’s not a piece of cake.

antiX utilities for processing PDF files (posted by ChPol)

[anticapitalistaForum Admin]

Well if an update breaks the init, users are F****d

Philosophers have interpreted the world in many ways; the point is to change it.

antiX with runit - leaner and meaner.

[Brian MasinickModerator]

anticapitalista wrote:

This is for antiX-26 trixie users, not testing or sid.

Should init(s) get updates?
There are updates to sysVinit and s6.
I successfully run both updated packages on my frugal trixie installs (both arches)
and the latest s6 on my antiX-23 installed system.

I’m in favor of inits, and for that matter, any installed packages and dependencies to get updates; the only exception I can think of would be broken packages held back until they get fixed.

--
Brian Masinick
Alternate "Search B":
This search page

[andfreeMember]

anticapitalista wrote:

if an update breaks the init, users are F****d

I think that this shouldn’t happen to users who haven’t chosen consciously to take such a risk.

antiX utilities for processing PDF files (posted by ChPol)

[anticapitalistaForum Admin]

andfree wrote:

anticapitalista wrote:

if an update breaks the init, users are F****d

I think that this shouldn’t happen to users who haven’t chosen consciously to take such a risk.

Users will have no (little) choice since updates will be automatic

Philosophers have interpreted the world in many ways; the point is to change it.

antiX with runit - leaner and meaner.

[andfreeMember]

Brian Masinick wrote:

the only exception I can think of would be broken packages held back until they get fixed.

Would the broken packages be identified as such early enough so that the users are not F****d?

antiX utilities for processing PDF files (posted by ChPol)

[christopheModerator]

anticapitalista wrote:

Users will have no (little) choice since updates will be automatic

I think the safety approach is best (ie., no upgrades).

A question I have, though, is why an update to the init? Is it typically to patch a flaw or to add new functionality?
It seems to me that if it’s working well, it shouldn’t need an upgrade mid-cycle.
And the risk of a “broken computer”, however potentially slight, should be avoided (IMO).

• This reply was modified 5 months, 4 weeks ago by christophe.

antiX on everything, since 2017

[anticapitalistaForum Admin]

andfree wrote:

Brian Masinick wrote:

the only exception I can think of would be broken packages held back until they get fixed.

Would the broken packages be identified as such early enough so that the users are not F****d?

Only if we have enough testers who are prepared to find out if an upgrade does actually break something.

The following scenarios would all need testing.

* running live/frugal booted into an init that is not going to be updated.
* running live/frugal booted into an init that is going to be updated.
* running installed booted into an init that is not going to be updated.
* running installed booted into an init that is going to be updated.
* checking iso-snaphot still works without issues in live mode
* checking iso-snaphot works without issues on an install.
* checking persistence/remastering is reliable when running live/frugal

Philosophers have interpreted the world in many ways; the point is to change it.

antiX with runit - leaner and meaner.

[anti-apXosMember]

I would prefer only security and bugfix updates a la Debian stable, but I_assume that would be extra work to apply only specific patches. So maybe only update them if there’s a critical fix, but then do a full update to the current master. Otherwise, no updates.

And if an update is coming, you could post on the forums to ask for extra testing.

• This reply was modified 5 months, 4 weeks ago by anti-apXos.

"--"

[Brian MasinickModerator]

I’ll try to test each of these scenarios as soon as I get a chance!

I will report back the status of my testing.

--
Brian Masinick
Alternate "Search B":
This search page

[andfreeMember]

anticapitalista wrote:

Only if we have enough testers who are prepared to find out if an upgrade does actually break something.

Will the testers have to check all the apps, so that they can make sure that nothing is broken?

antiX utilities for processing PDF files (posted by ChPol)

[olsztynMember]

anticapitalista wrote:

andfree wrote:
anticapitalista wrote:
if an update breaks the init, users are F****d

I think that this shouldn’t happen to users who haven’t chosen consciously to take such a risk.

Users will have no (little) choice since updates will be automatic

I do not see a problem here. Every antiX user should have a version control in place, so any unfortunate update can be backed out through restore or rollback…

Personally I do hope there will be many aggressive updates to inits as they evolve to be more and more mature. I am still not able to run dinit correctly after all and some other random issues…

Live antiX Boot Options (Previously posted by Xecure):
http://antixlinuxfan.miraheze.org/wiki/Table_of_antiX_Boot_Parameters

[grindstoneMember]

I think the multi-init was too ambitious for the available resources/timing — now that it’s out, I say treat it like the RC it might’ve been in a different universe where anti et al had unlimited time (but with the testing known since release).

If the init updates need to happen (probably do as it has been a heavy lift) AND there are not means/volunteers to get booting status coverage, then maybe some grand notice to remaster between updates (?) can be prominently recommended.

I’m in the worst possible hardware setup (laptop) right now, but, if it helps, I can help test in vbox if its worth anything. Just need get definitions of baseline configs to use/retain and objectives.

FWIW.

[anti-apXosMember]

Maybe @anticapitalista could clarify what packages this questions was about. I assumed it was to do with the main init system packages themselves, not associated service packages or other helpers used by the inits, which seems to be where most of the fine-tuning for init-diversity happens.

"--"

[Author]

Posts