Hi Gabriel,

On 2026-04-27 at 16:30+02:00, Gabriel Wicki wrote:
> Please let me know about any minor and major issues
> you still see with the proposal

I'm sorry I did not respond to your reply sooner, so here it goes:

On 2026-04-03 at 15:45+02:00, Gabriel Wicki wrote:
> Nguyễn Gia Phong writes:
> > I do like being reached out to, I just do not want it to become
> > a responsibility across all my communication channels.
> > May I suggest narrowing this down to communication channels registered
> > for the project?
>
> That is what is intended.  Nobody should be reachable for the project on
> channels they did not themselves add (or approve to be added) to
> `etc/teams.scm`.

Thanks, this section currently reads:

> Team members are expected to be reachable and respond to communications
> both from within as well as from outside of the project by email,
> and for code-related teams Codeberg, through their team's label.

I suggest rephrasing the latter part something along the line of

> communications through channels registered in `etc/teams.scm`,
> which at the time of writing includes email and Codeberg.

Alone, _email_ could entails any other address and _communications
from outside of the project_ still reads to me like stuff additional
to the later listed email and Codeberg.  Since the social scope
of the project is not clearly defined, I think we can drop
the within & outside part (which just means everyone anyway).

> Active committers should be informed about ongoing discussions
> with a frequency of at least 7 days.

I do not think this is possible even if we only count guix-devel
for the discussions.  Since it's to

> ensures that all active committers take note and can respect
> imposed restrictions that are taken by other committers
> (like a push freeze to the `master` branch prior to a release).

I suggest requiring commiters to be informed of the outcome
of such discussions, e.g. guix-info.  There are decisions
one may have no interest in discussing and just go with
whatever that is decided.

On 2026-04-03 at 15:45+02:00, Gabriel Wicki wrote:
> Nguyễn Gia Phong writes:
> > I prefer to not have this classification as the difference
> > between passive and inactive is unclear, but rather a process
> > for committers to be unreachable.
>
> I've tried to clarify this distinction more explicitly with the upcoming
> change since it has spurred quite some confusion and disagreement.
>
> Nguyễn Gia Phong writes:
> > I suggest committers to announce their unavailability period in the
> > teams file, either directly or through other commiters (e.g. their
> > absence is due to being AFK).  This information is invaluable to other
> > contributors so they know who to ping.
>
> To me "passive" committers are not people on a prolonged (think:
> multiple months) leave, but rather people temporarily unavailable to do
> some work.  I don't think it is necessary to announce single-day or
> week-long absences publicly, but I can see how teamwork could use
> arrangements with respect to longer absences.  I'd leave that up to the
> teams.

Currently, the definition of passive committers is only used
to require them to catch up with the state of affairs
before transforming into active ones.  How about

> before pushing changes to protected branches, commiters must
> be updated with the relevant decisions that are announced in [...]

for the entire reachability section?  As a team members,
they need to be reachable via registered communication means anyway.

On 2026-04-03 at 15:45+02:00, Gabriel Wicki wrote:
> Nguyễn Gia Phong writes:
> > >  - The commiter's OpenPGP fingerprint is added to the
> > >    .guix-authorizations file of the branch(es) you will commit to.
> >
> > Is this leaking the idea of explicit commit access to each branch?
> > I think that is a new thing by itself and not described in this GCD.
> > Either way, I think this risks merging commit rights accidentally,
> > e.g. from a staging branch to master.
>
> Currently this would mean only adding the fingerprint to the `master`
> branch.  I can gladly remove this bullet-point if it is too confusing or
> does not adequately reflect our current handling of the situation.

Please do, I support the idea of explicit branch access,
but its execution needs to be discussed upon (later).

Best wishes,
Phong

Attachment: signature.asc
Description: PGP signature

Reply via email to