Prev: [PATCHv2] regulator: Get rid of lockdep warning
Next: rfkill sysfs ABI
From: Linus Torvalds on 12 Mar 2010 13:20


On Wed, 10 Mar 2010, James Bottomley wrote:
>
> OK, I know this is a bit late, but no-one told me the merge window was
> going to close early, but even if they had I'd have been unable to do
> anything about it, since I've been in transit from India all this week
> so far (don't ask ...).

I'm not going to pull.

The whole "a few days early" was literally aimed at you and some similar
people who _always_ send the pull request late, literally the last days.

I'm fed up with it.

The damn thing should have been ready when the merge window _opened_, not
just before closing. And now there _is_ a tree waiting for me for the next
merge window. Goodie.

Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: James Bottomley on 12 Mar 2010 13:30
On Fri, 2010-03-12 at 10:10 -0800, Linus Torvalds wrote:
>
> On Wed, 10 Mar 2010, James Bottomley wrote:
> >
> > OK, I know this is a bit late, but no-one told me the merge window was
> > going to close early, but even if they had I'd have been unable to do
> > anything about it, since I've been in transit from India all this week
> > so far (don't ask ...).
>
> I'm not going to pull.
>
> The whole "a few days early" was literally aimed at you and some similar
> people who _always_ send the pull request late, literally the last days.
>
> I'm fed up with it.
>
> The damn thing should have been ready when the merge window _opened_, not
> just before closing. And now there _is_ a tree waiting for me for the next
> merge window. Goodie.

I do always have a tree ready in the first days of the merge window. It
was, in fact, here:

commit 654451748b779b28077d9058442d0f354251870d
Merge: 64d497f 77c9cfc
Author: Linus Torvalds <torvalds(a)linux-foundation.org>
Date: Fri Feb 26 16:55:27 2010 -0800

Merge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6

The slight problem is that I accumulate updates (like almost everyone
else) from the few days before the merge window opened up until about
the middle of the second week. That's what goes into the final pull,
plus assorted bug fixes that showed up during the window.

I'm not objecting to a shorter merge window, just the unpredictability
of all of this. The reason I delay the second pull until the last days
of the merge window is so that the proposed changes get the maximum
amount of shake down in linux-next ... and that usually means I close
down the inclusion window about three days before the pull request goes
out.

Personally, I do like the two week merge window, because it shakes out
contributors to me who should have got their act together but didn't.
If you don't like the way this works, fine ... I just need to know what
the parameters are to adjust accordingly, so what is it now? A 10 day
merge window?

James


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Linus Torvalds on 12 Mar 2010 13:40


On Fri, 12 Mar 2010, James Bottomley wrote:
>
> I'm not objecting to a shorter merge window, just the unpredictability
> of all of this.

If you read the 2.6.33 release notes, the unpredictability WAS THE WHOLE
F*CKING POINT!

I'm tired of people being able to predict exactly when the merge window
closes, and then using that to make the last days of the merge window be
hell on me, and sending all the merges within about 24 hours of the
closing.

So I'm not doing it.

If you had the major changes ready when the merge window opened, you
should damn well have sent it in _then_.

End of story. I've told people before. The merge window is for _merging_,
not for doing development. If you send me something the last day, then
there is no "window" any more. And It is _really_ annoying to have fifty
pull requests on the last day. I'm not going to take it any more.

And the thing is, all your explanations make things WORSE. You're
basically saying that you're working exactly the way that I'm not going to
accept.

So you're now going to get to wait six or seven weeks, because I refuse to
work the way you have worked me for the last two years. I gave people a
heads-up, and I probably shouldn't have. Some people noticed. And other
people didn't even care, because they never gamed the system in the first
place.

And some people - like you - both gamed the system, and didn't read the
release notes, and now you complain. Tough.

And the fact that you complain about the very fact that you were gaming
the system and now you can't because it's not predictable enough to game
is what just proves my f*cking point!

Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Linus Torvalds on 12 Mar 2010 13:50


On Fri, 12 Mar 2010, Linus Torvalds wrote:
>
> And the fact that you complain about the very fact that you were gaming
> the system and now you can't because it's not predictable enough to game
> is what just proves my f*cking point!

Btw, just to clarify for next time: I'm not even going to mention in the
release notes how long the merge window is going to be.

Maybe I'll say "that's enough" after just one week, because the good
people did what they do, and sent me a lot of merges early, there were no
problems, and I might as well cut an -rc1 early since I have enough and
I'd like to occasionally get a _shorter_ release cycle not just because
the merge window was shorter, but because hopefully merging less might
result in an easier stabilization cycle too.

Or maybe I'll be busy for a few days, and extend it past the two weeks
just because I was in the middle of another flame war and needed a few
extra days to catch up with real work.

Who knows? The point being that being predictable has caused problems.
People have decided that they have an extra two weeks, and that the merge
window isn't a window at all, it's just a day when -rc1 gets done.

Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Ray Lee on 13 Mar 2010 14:40
On Fri, Mar 12, 2010 at 10:27 AM, James Bottomley wrote:
> I do always have a tree ready in the first days of the merge window.  It
> was, in fact, here:

[...]

> The slight problem is that I accumulate updates (like almost everyone
> else) from the few days before the merge window opened up until about
> the middle of the second week.  That's what goes into the final pull,
> plus assorted bug fixes that showed up during the window.

You appear to being using git as a sort of baroque CVS. Fixes that
arrive after the beginning of the merge window should be put into a
separate branch other than the one ready for submission. Features that
arrive after the beginning of the merge window should be placed into
yet another branch destined for the next merge window.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
 |  Next  |  Last
Pages: 1 2
Prev: [PATCHv2] regulator: Get rid of lockdep warning
Next: rfkill sysfs ABI