From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751301Ab2AYW3Q (ORCPT ); Wed, 25 Jan 2012 17:29:16 -0500 Received: from ogre.sisk.pl ([217.79.144.158]:50345 "EHLO ogre.sisk.pl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750772Ab2AYW3O (ORCPT ); Wed, 25 Jan 2012 17:29:14 -0500 From: "Rafael J. Wysocki" To: Paul Gortmaker Subject: Re: [Announce] linux-pm tree changes Date: Wed, 25 Jan 2012 23:32:48 +0100 User-Agent: KMail/1.13.6 (Linux/3.3.0-rc1+; KDE/4.6.0; x86_64; ; ) Cc: Linux PM list , LKML References: <201201242356.04244.rjw@sisk.pl> In-Reply-To: MIME-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Message-Id: <201201252332.48993.rjw@sisk.pl> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wednesday, January 25, 2012, Paul Gortmaker wrote: > On Tue, Jan 24, 2012 at 5:56 PM, Rafael J. Wysocki wrote: > > Hi, > > > > I'd like you to know some workflow changes regarding the linux-pm tree. > > > > Since it's now possible to create signed tags for Linus to pull from, I won't > > need the 'pm-for-linus' branch any more and it's gone. The 'pm-fixes' branch > > is gone too and replaced by the 'fixes' branch (which currently contains the > > same material as master). > > > > The idea is that the main development is going to happen in the 'master' > > branch, which is not going to be rebased and the Linus' tree will be merged > > into it from time to time (most of the time when the 'master' branch can be > > fast-forwarded to it). The 'fixes' branch will be used for pushing urgent > > Not sure what you have in mind for "from time to time", Something like once in a development cycle (right after a merge window), but there may be exceptions. > but random merges > without a clear reason (i.e. "merge Linus to get access to new files X Y Z") > are probably going to get you in hot water. I know. > Also, once you stack on your 1st PM commit onto your master, none of > your future merges of Linus' tree will be fast forward. Until Linus pulls from my tree. Then I can do a fast forward and start over. > Here is an > example, where I simulate a local change, followed by "time to time" > merges of Linus tree. Note each creates a merge, not just the 1st one. > > ----------------------------- > linux-head$git checkout -b crap v3.1 > Checking out files: 100% (17551/17551), done. > Switched to a new branch 'crap' > linux-head$echo foo > foo ; git add foo ; git commit -m '1st pm commit' > [crap 1675da3] 1st pm commit > 1 files changed, 1 insertions(+), 0 deletions(-) > create mode 100644 foo > linux-head$git merge v3.2-rc1 2>&1 |grep 'Merge\|Fast' > Merge made by recursive. > linux-head$git merge v3.2-rc2 2>&1 |grep 'Merge\|Fast' > Merge made by recursive. > linux-head$git merge v3.2-rc3 2>&1 |grep 'Merge\|Fast' > Merge made by recursive. > linux-head$git show -m |head -n5 > commit 3cbd041875960361ffde72612afadc6c68e94654 (from > 6f6028bdbbc63dc71b408da8976f154a2e6a70ee) > Merge: 6f6028b caca6a0 > Author: Paul Gortmaker > Date: Wed Jan 25 13:45:55 2012 -0500 > linux-head$ > ---------------- > > Having lots of these kinds of pointless merge commits are > what will most likely get you in hot water when you ask for > it to be pulled. > > Maybe I'm just misinterpreting what your plan is. Please see above. :-) Thanks, Rafael