Re: "If you're still seeing problems, please reopen." [Was: bug#25148:]

["João Távora" <joaotavora@gmail.com>]

On Thu, Nov 21, 2019 at 6:26 PM Eli Zaretskii <eliz@gnu.org> wrote:

> > > which will make it be hosted on
> > > the same machine where our upstream repository lives.  And then who's
> > > to say that such branches pushed into our GitLab are not part of
> > > Emacs, like all the scratch branches you and others push now?
> >
> > Is that question really important?
>
> Yes.  Because by hosting code that violates some copyright we could
> make GNU liable.  Or if someone, by mistake or malice, pushes code
> that is against the GNU policies, we could make it easy for enemies of
> GNU to attack and discredit GNU.  Etc. etc.
> > And why does it not apply to patches as potentially absurd or as
> > harmful as those branches that are sitting in email bodies of the
> > debbugs bug tracker and the archives of lists.gnu.org?
> Because an email message is from someone who offers us code, wheres
> code on our server is already on our server, in our repository, even
> if it's a fork.  Thus, people who don't understand these
> technicalities could easily be convinced that we "endorsed" that code.

Regarding perception, a lot of people that have "absurd",
"ilegal", "malicious", etc. forks of other projects on GitHub,
but I don't have reasons to believe the project's canonical
repos are worried about that perception.

But indeed, GitHub is different than gitlab.mycompany.com so
maybe Richard and the FSF legal experts could
look into that.

> > But what does that risk amount to, i.e. what are
> > the consequences of someone making that mistake?
>
> In our present chaotic world, the consequences might be dire.

OK. A bit open-ended, but I really don't have the arguments
to dispute that.

> And I really don't understand the need for trying to do it on
> Savannah, when savannah.nongnu is a good compromise that solves this
> problem.  Why should anyone care on what server is this tracker
> hosted?

Oh, I didn't propose that, you did (I think?). I just assumed it
was a requirement. I personally don't care what server the
canonical fork of Emacs is in.

João