On Fri, 18 May 2018 at 15:30:27 -0700
Rick Moen <rick@???> wrote:

> Quoting Alessandro Selli (alessandroselli@???):
>
>> Adam Borowski's point is that if you have a single person
>> administering the distro's infrastructure it does not matter if the
>> infrastructure is an autonomous one or if it's hosted on GitHub:
>
> So, Don't Do That, Then.

Whom shouldn't do that, the distro developer or the distro's admin?
Typically developers and contributors do not have the master repositories'
"keys", so it does not matter if those keys open up a personal repo, or a
GitLab or a GitHub one. If the admin has the only set of keys and suddently
disappears, repos are inaccessible no matter where they are. "Don't Do
That", sure, but people do that, and it's their decision over which you
have no control.

> {headdesk}
>
> I really don't think it's that difficult to understand that avoiding
> outsourcing in no way precludes appropriate fallbacks and measures to
> eliminate SPoFs. I'm frankly quite puzzled that my mentioning (as an
> example) GitLab elicited the comment 'This wouldn't have helped
> [because] you need redundancy' -- when I nowhere suggested eschewing
> redundancy and when that open source project has a mountain of
> documentation on that very subject. And I'm puzzled a second time to
> see you ignore my having just pointed that out, as if I hadn't.

You need redundancy in repository's admins, not on infrastructure.

> And redundant infrastructure (obviously) facilitates redundant
> administration and oversight.

Did you miss in this case "redundant administration" is exactly what is
amiss? Was the main repo on GitLab or WhateverHub nothing would have
changed: develpers whould still be unable to access it.

> Why is this difficult? I don't think it is.


I think you misunderstood what was meant by "redundant".


Alessandro