[Nix-dev] Policy for updates in 14.04

[Lluís Batlle i Rossell]

On Sun, Aug 31, 2014 at 06:31:50PM +0200, Peter Simons wrote:
> Hi Lluís,
> 
>  > Almost any software update will contain a bunch of bugfixes and only
>  > sometimes new features.
> 
> can you point me to an empirical study that supports this thesis? 

I mean the impression I got from reading release notes of different programs.
For example:
https://gitweb.torproject.org/tor.git?a=blob_plain;hb=release-0.2.4;f=ReleaseNotes

>  > And for lots of nixpkgs software, having the update in master,
>  > doesn't guarantee that it is much tested once it becomes stable.
> 
> I'm not sure whether I understand your point correctly. How does this
> relate to the policy for the stable release branches? Can you maybe
> re-phrase this argument?

I mean that, for some software, the stable branch does not run into much more
tests than master. And at some point, master becomes the new stable. There is no
gathering of signatures from maintainers testifying that "it works" or similar.

My argument is that I don't think that having very strict policies of update
will not make the stable branch relevantly safer.

>  > I'd prefer to have a less strict rule about updating the stable
>  > branch, and act based on problems arisen. 
> 
> Well, this is exactly the kind of policy I *don't* want on the stable
> branch. In my humble opinion, it should be extremely safe to run
> "nixos-rebuild --upgrade switch" on a stable branch at any point in
> time. I don't want patches in there that may break my system. If I'm
> okay with running into bugs that may be fixed later, then I'll use
> 'master'.

Sure, I agree with those patches. That's why I made a distinction about "system
packages". If I update a bittorrent client in the release branch, having tested
it, should do no harm. And if there are cases of harm inflicted, we should talk
about them, not about hypotetical. I mean that maybe that despite not being a
full agreement about what can be committed to the stable branch or not, the
result is that the common sense of most committers makes it stable enough.

>  > Depending on the strictness, few are going to use the stable branch.
>  > For me, I choose more the stable branch for much less rebuilds. No
>  > mesa updates, no X updates, no qt updates, no gtk udpates, etc...
> 
> Do we have any data that suggests how popular the 14.04 branch is
> compared to master? Maybe the Hydra log files can shed some light on
> that question?

Let's see if anyone provides any figures. As for me, most of my computers run
the stable branch, but as I say, I commit there updates for some software I care
of, and that I think will do no harm.

Regards,
Lluís.