Re: Please create the email alias do-not-apply-to-stable@xxxxxxxxxx -> /dev/null

From: Mauro Carvalho Chehab
Date: Mon Apr 22 2024 - 18:16:02 EST


Em Tue, 23 Apr 2024 00:04:01 +0200
Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> escreveu:

> On Mon, Apr 22, 2024 at 10:46:37PM +0100, Mauro Carvalho Chehab wrote:
> > Em Mon, 22 Apr 2024 15:25:18 -0400
> > Konstantin Ryabitsev <konstantin@xxxxxxxxxxxxxxxxxxx> escreveu:
> >
> > > On Mon, Apr 22, 2024 at 05:49:29PM +0200, Thorsten Leemhuis wrote:
> > > > @Greg, BTW: should this be stable+noautosel@xxxxxxxxxx or have a
> > > > 'vger.'
> > >
> > > No vger, just stable+whatever@xxxxxxxxxx.
> > >
> > > > in it, e.g. stable+noautosel@xxxxxxxxxxxxxxx? I assume without 'vger.'
> > > > is fine, just wanted to be sure, as
> > > > Documentation/process/stable-kernel-rules.rst in all other cases
> > > > specifies stable@xxxxxxxxxxxxxxx, so people are likely to get confused.
> > > > :-/ #sigh
> > >
> > > These serve two different purposes:
> > >
> > > stable@xxxxxxxxxx (goes into devnull)
> > > stable@xxxxxxxxxxxxxxx (actual mailing list)
> > >
> > > Confusion happens all the time, unfortunately.
> >
> > Yeah, I did already used stable@xxxxxxxxxx a few times in the
> > past.
> >
> > IMO, the best would be either for stable to also accept it or for
> > kernel.org mail server to return an error message (only to the
> > submitter) warning about the invalid address, eventually with a
> > hint message pointing to the correct value.
>
> stable@xxxxxxxxxx is there to route to /dev/null on purpose so that
> developers/maintainers who only want their patches to get picked up when
> they hit Linus's tree, will have happen and not notify anyone else.
> This is especially good when dealing with security-related things as we
> have had MANY people accidentally leak patches way too early by having
> cc: stable@xxxxxxxxxxxxxxx in their signed-off-by areas, and forgetting
> to tell git send-email to suppress cc: when sending them out for
> internal review.

Nice! didn't know about that. On a quick check, the only place at
documentation mentioning it without vger is at checkpatch.rst.

Perhaps it would make sense to document that as well.

> Having that bounce would just be noisy for the developers involved.
>
> thanks,
>
> greg k-h