Re: Linux 4.20-rc6
From: Andy Shevchenko
Date: Mon Dec 10 2018 - 13:39:49 EST
(Oops, forgot to Cc Darren, fixing)
On Mon, Dec 10, 2018 at 7:46 PM Linus Torvalds
> On Mon, Dec 10, 2018 at 1:52 AM Andy Shevchenko
> <andy.shevchenko@xxxxxxxxx> wrote:
> > Linus, due to my vacation schedule and possibility that Darren might
> > be also busy with some stuff, would be okay to send PR this week, for
> > example?
> Oh, absolutely.
Thanks for elaborated answer!
> And I want to make it clear that this is not something you ever need
> to even ask about, or make some "special case". It is absolutely
> always true that any submaintainer can send pull requests for the next
> merge window early.
> To avoid any confusion, I do ask for a couple of simple things if you
> send a pull request before the merge window opens:
> - make it very clear and _explicit_ in the pull request that this is
> for the next merge window.
> In other words, don't make me unnecessarily upset about "this is
> much too late in the rc series to send something this big" by making
> it a small note at the end.
> This could still be something lik [GIT PULL] in the subject line,
> but on the very first line say "For the next merge window" or
> something. Don't make it subtle or something I might not notice before
> I start getting upset at how big a pull request you're sending me
> during rc6+.
> - don't use it as an excuse to send two or more bigger merge window
> pulls. An early pull is not an excuse to then have a second big pull
> with other changes.
> Second (and third) pulls are obviously fine, but they should either
> be starting the calming-down period, or be about other things entirely
> (some groups send 5-10 pull requests that are independent).
> I don't want to get the feeling that there is a "this is partial
> work, there will be more" pattern.
> Anyway, people have done this in the past, and I don't mind it at all.
> Send pull requests early if they are all done and ready.
> You don't even need to have a vacation excuse, you can just say "I'm
> done, this is already in linux-next and it's passed testing etc, it's
> not some unbaked thing that I just want to get rid of early"
With Best Regards,