git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* GIT and the current -stable
@ 2007-04-14  5:36 Rene Herman
  2007-04-14  6:24 ` Junio C Hamano
                   ` (2 more replies)
  0 siblings, 3 replies; 16+ messages in thread
From: Rene Herman @ 2007-04-14  5:36 UTC (permalink / raw)
  To: Linux Kernel; +Cc: Greg KH, git

Good day.

Stumbling around with git here. I'd like to use git to efficiently track the 
  current -stable as well as -current. Say, my local tree is a clone of 
Linus current:

git clone \ 
git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git local

I then branch off a 2.6.20 branch:

cd local
git checkout -b v2.6.20 v2.6.20

to now update to the current -stable I could do:

git pull \ 
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.20.y.git

each time that a new -stable is released. Rather though, I'd like a simple 
"git pull" to do this while on this branch while a "git pull" while back on 
the master branch pulls from the originally cloned Linus repo again.

Is this possible? Do I want it to be? Comments like "work like this instead" 
welcome as well; figuring out what the heck it is that I want from git seems 
to be one of the most difficult questions to answer...

Currently using git 1.5.0.4.

Rene.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14  5:36 GIT and the current -stable Rene Herman
@ 2007-04-14  6:24 ` Junio C Hamano
  2007-04-14  7:13   ` Rene Herman
                     ` (2 more replies)
  2007-04-14  8:34 ` Chris Wright
  2007-04-14 11:56 ` Uwe Kleine-König
  2 siblings, 3 replies; 16+ messages in thread
From: Junio C Hamano @ 2007-04-14  6:24 UTC (permalink / raw)
  To: Rene Herman; +Cc: Linux Kernel, Greg KH, git

Rene Herman <rene.herman@gmail.com> writes:

> Stumbling around with git here. I'd like to use git to efficiently
> track the current -stable as well as -current. Say, my local tree is a
> clone of Linus current:
>
> git clone \
> git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
> local
>
> I then branch off a 2.6.20 branch:
>
> cd local
> git checkout -b v2.6.20 v2.6.20
>
> to now update to the current -stable I could do:
>
> git pull \
> git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.20.y.git

I think adding these lines to .git/config would do the trick,
after you have done the "checkout -b v2.6.20 v2.6.20" step:

[branch "v2.6.20"]
	remote = stable
	merge = refs/heads/master
[remote "stable"]
	url = git://git.kernel.org/.../stable/linux-2.6.20.y.git
	fetch = refs/heads/master

provided if stable team forks v2.6.20.y history off of Linus's
v2.6.20.

With the above configuration, anytime you say "git pull" while
on your v2.6.20 branch will fetch from "stable" and merge their
'master' branch in your current branch (i.e. v2.6.20 branch).

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14  6:24 ` Junio C Hamano
@ 2007-04-14  7:13   ` Rene Herman
  2007-04-14  7:21   ` Greg KH
  2007-04-14 17:34   ` Linus Torvalds
  2 siblings, 0 replies; 16+ messages in thread
From: Rene Herman @ 2007-04-14  7:13 UTC (permalink / raw)
  To: Junio C Hamano; +Cc: Linux Kernel, Greg KH, git

On 04/14/2007 08:24 AM, Junio C Hamano wrote:

> I think adding these lines to .git/config would do the trick,
> after you have done the "checkout -b v2.6.20 v2.6.20" step:
> 
> [branch "v2.6.20"]
> 	remote = stable
> 	merge = refs/heads/master
> [remote "stable"]
> 	url = git://git.kernel.org/.../stable/linux-2.6.20.y.git
> 	fetch = refs/heads/master
> 
> provided if stable team forks v2.6.20.y history off of Linus's
> v2.6.20.
> 
> With the above configuration, anytime you say "git pull" while
> on your v2.6.20 branch will fetch from "stable" and merge their
> 'master' branch in your current branch (i.e. v2.6.20 branch).

Yes, this does seem to work, thanks. Was thrown of a bit by having named the 
branch "v2.6.20". GIT and I disagree what it is that I want to happen when I 
say "git checkout v2.6.20" if v2.6.20 is also a tag on master.

The pull behaviour does not follow further branches:

rene@7ixe4:~/src/linux/local$ git branch
* 2.6.20
   master
rene@7ixe4:~/src/linux/local$ git checkout -b 7ixe4
Switched to a new branch "7ixe4"
rene@7ixe4:~/src/linux/local$ git pull
Warning: No merge candidate found because value of config option
          "branch.7ixe4.merge" does not match any remote branch fetched.
No changes.

This might in practice not be all bad in fact, and I suppose I understand 
how to "fix" it along the same lines as above.

But as happens very frequently with GIT, I get the feeling that I just don't 
understand how it's all intended to be used. It seems that what I wanted 
above is not standard? What would be expected use of the stable GIT repo? 
Just cloning that outright into another repo?

A "GIT WHYTO" from someone with the oversight would be very useful...

Rene.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14  6:24 ` Junio C Hamano
  2007-04-14  7:13   ` Rene Herman
@ 2007-04-14  7:21   ` Greg KH
  2007-04-14 17:34   ` Linus Torvalds
  2 siblings, 0 replies; 16+ messages in thread
From: Greg KH @ 2007-04-14  7:21 UTC (permalink / raw)
  To: Junio C Hamano; +Cc: Rene Herman, Linux Kernel, git

On Fri, Apr 13, 2007 at 11:24:19PM -0700, Junio C Hamano wrote:
> Rene Herman <rene.herman@gmail.com> writes:
> 
> > Stumbling around with git here. I'd like to use git to efficiently
> > track the current -stable as well as -current. Say, my local tree is a
> > clone of Linus current:
> >
> > git clone \
> > git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
> > local
> >
> > I then branch off a 2.6.20 branch:
> >
> > cd local
> > git checkout -b v2.6.20 v2.6.20
> >
> > to now update to the current -stable I could do:
> >
> > git pull \
> > git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.20.y.git
> 
> I think adding these lines to .git/config would do the trick,
> after you have done the "checkout -b v2.6.20 v2.6.20" step:
> 
> [branch "v2.6.20"]
> 	remote = stable
> 	merge = refs/heads/master
> [remote "stable"]
> 	url = git://git.kernel.org/.../stable/linux-2.6.20.y.git
> 	fetch = refs/heads/master
> 
> provided if stable team forks v2.6.20.y history off of Linus's
> v2.6.20.

Yes, we do branch off Linus's tree for each of the different .y
releases.

Some people have asked us to keep all of the different .y branches in
the same git tree, but for now we have been using different repositories
on kernel.org to keep things simpler for the majority of people who do
not fully understand how to use the branching features of git (myself
included in that category at times, I've messed up merging to different
branches too many times...)

> With the above configuration, anytime you say "git pull" while
> on your v2.6.20 branch will fetch from "stable" and merge their
> 'master' branch in your current branch (i.e. v2.6.20 branch).

That's really cool, thanks for pointing out how to do it.

thanks,

greg k-h

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14  5:36 GIT and the current -stable Rene Herman
  2007-04-14  6:24 ` Junio C Hamano
@ 2007-04-14  8:34 ` Chris Wright
  2007-04-14  8:54   ` Rene Herman
  2007-04-14 15:15   ` Brian Gernhardt
  2007-04-14 11:56 ` Uwe Kleine-König
  2 siblings, 2 replies; 16+ messages in thread
From: Chris Wright @ 2007-04-14  8:34 UTC (permalink / raw)
  To: Rene Herman; +Cc: Linux Kernel, Greg KH, git

* Rene Herman (rene.herman@gmail.com) wrote:
> Stumbling around with git here. I'd like to use git to efficiently track 
> the current -stable as well as -current. Say, my local tree is a clone of 
> Linus current:
> 
> git clone \ 
> git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git local
> 
> I then branch off a 2.6.20 branch:
> 
> cd local
> git checkout -b v2.6.20 v2.6.20
> 
> to now update to the current -stable I could do:
> 
> git pull \ 
> git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.20.y.git

I've already put a tree like this up on kernel.org.  The master branch
is Linus' tree, and there's branches for each of the stable releases
called linux-2.6.[12-20].y (I didn't add 2.6.11.y).

http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6-stable.git;a=summary

> each time that a new -stable is released. Rather though, I'd like a simple 
> "git pull" to do this while on this branch while a "git pull" while back on 
> the master branch pulls from the originally cloned Linus repo again.

You have to be careful with pull.  It will always want to merge onto your
current branch.

thanks,
-chris

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14  8:34 ` Chris Wright
@ 2007-04-14  8:54   ` Rene Herman
  2007-04-14 23:13     ` Rene Herman
  2007-04-14 15:15   ` Brian Gernhardt
  1 sibling, 1 reply; 16+ messages in thread
From: Rene Herman @ 2007-04-14  8:54 UTC (permalink / raw)
  To: Chris Wright; +Cc: Linux Kernel, Greg KH, git

On 04/14/2007 10:34 AM, Chris Wright wrote:

> I've already put a tree like this up on kernel.org.  The master branch
> is Linus' tree, and there's branches for each of the stable releases
> called linux-2.6.[12-20].y (I didn't add 2.6.11.y).
> 
> http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6-stable.git;a=summary

I see, thank you; that sounds like a good "master" repo to clone then.

Rene.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14  5:36 GIT and the current -stable Rene Herman
  2007-04-14  6:24 ` Junio C Hamano
  2007-04-14  8:34 ` Chris Wright
@ 2007-04-14 11:56 ` Uwe Kleine-König
  2 siblings, 0 replies; 16+ messages in thread
From: Uwe Kleine-König @ 2007-04-14 11:56 UTC (permalink / raw)
  To: Rene Herman; +Cc: Linux Kernel, Greg KH, git

[-- Attachment #1: Type: text/plain, Size: 1314 bytes --]

Hello Rene,

Rene Herman wrote:
> each time that a new -stable is released. Rather though, I'd like a simple 
> "git pull" to do this while on this branch while a "git pull" while back on 
> the master branch pulls from the originally cloned Linus repo again.
> 
> Is this possible? Do I want it to be? Comments like "work like this 
> instead" welcome as well; figuring out what the heck it is that I want from 
> git seems to be one of the most difficult questions to answer...
Chris' mail made that somehow superfluous but anyhow.  I have a Python
script to update my Linux tree.  It's attached.

It automatically adds
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.x.y as a
remote as soon as v2.6.x is seen in Linus' repo and then fetches 2.6.16,
2.6.(last - 1) and 2.6.last.  If you want to fetch all, change 

	for stable in (16, last_stable - 1, last_stable):

in

	for stable in xrange(16, last_stable + 1):

The location of the repo is hardcoded to $HOME/gsrc/linux-2.6, but I
think everyone with a bit of intuition should be able to change that ...

So it's not a "git pull" for me, but "update-linux".

Best regards
Uwe

-- 
Uwe Kleine-König

main(){char*a="main(){char*a=%c%s%c;printf(a,34,a,34%c";printf(a,34,a,34
,10);a=",10);a=%c%s%c;printf(a,34,a,34,10);}%c";printf(a,34,a,34,10);}

[-- Attachment #2: update-linux --]
[-- Type: text/plain, Size: 1832 bytes --]

#! /usr/bin/env python

import os, re, subprocess, sys

silent = False

git_dir = os.path.join(os.getenv('HOME'), 'gsrc', 'linux-2.6', '.git')
os.putenv('GIT_DIR', git_dir)

devnull = open('/dev/null', 'w', 0)
if silent:
  subprocess_stdout = devnull
else:
  subprocess_stdout = None

if not silent:
  print 'fetch origin'
subprocess.call(['git-fetch', 'origin'], stdout=subprocess_stdout)

sp_describe = subprocess.Popen(['git-describe', 'origin'], stdout=subprocess.PIPE, stderr=subprocess.PIPE)
sp_describe_stdin, sp_describe_stderr = sp_describe.communicate()

re_linuxver = re.compile('v2\.6\.(?P<minor>\d+)(?P<stable>\.\d+)?(?P<rc>-rc\d+)?(?P<git>-g[0-9a-fA-F]+)?')
mo = re_linuxver.match(sp_describe_stdin)

if not mo:
  print 'cannot parse version of origin\'s HEAD'
  sys.exit(1)

modict = mo.groupdict()

if modict['rc'] and not modict['stable']:
  last_stable = int(modict['minor']) - 1
else:
  last_stable = int(modict['minor'])

for stable in (16, last_stable - 1, last_stable):
  if subprocess.call(['git', 'repo-config', 'remote.v2.6.%d.y.url' % stable], stdout=devnull) != 0:
    subprocess.call(['git', 'repo-config', 'remote.v2.6.%d.y.url' % stable, 'git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.%d.y.git' % stable])
    subprocess.call(['git', 'repo-config', 'remote.v2.6.%d.y.fetch' % stable, 'refs/heads/master:refs/remotes/stable/v2.6.%d.y' % stable])
    subprocess.call(['git', 'repo-config', 'branch.v2.6.%d.y.remote' % stable, 'v2.6.%d.y' % stable])
    subprocess.call(['git', 'repo-config', 'branch.v2.6.%d.y.remote' % stable, 'v2.6.%d.y' % stable])
    subprocess.call(['git', 'repo-config', 'branch.v2.6.%d.y.merge' % stable, 'refs/heads/master'])

  if not silent:
    print 'fetch v2.6.%d.y' % stable
  subprocess.call(['git-fetch', 'v2.6.%d.y' % stable], stdout=subprocess_stdout)

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14  8:34 ` Chris Wright
  2007-04-14  8:54   ` Rene Herman
@ 2007-04-14 15:15   ` Brian Gernhardt
  2007-04-14 17:19     ` Chris Wright
  1 sibling, 1 reply; 16+ messages in thread
From: Brian Gernhardt @ 2007-04-14 15:15 UTC (permalink / raw)
  To: Chris Wright; +Cc: Rene Herman, Linux Kernel, Greg KH, git


On Apr 14, 2007, at 4:34 AM, Chris Wright wrote:

> I've already put a tree like this up on kernel.org.  The master branch
> is Linus' tree, and there's branches for each of the stable releases
> called linux-2.6.[12-20].y (I didn't add 2.6.11.y).
>
> http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6- 
> stable.git;a=summary

Is HEAD for that repo the most recent stable branch, or (as gitweb  
makes it look) Linus's head.  I'd expect a "-stable" repo to point at  
the most recent stable commit, not the most recent development  
commit.  And I'd also expect gitweb's summary page to show the  
shortlog for HEAd.  One of my assumptions are being broken and I  
don't like it.  It leaves me all confused...

(I'd just clone your repo and look for myself, but I'm still pulling  
Linus's tree for the first time and I don't have a fast connection.)

~~ Brian

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14 15:15   ` Brian Gernhardt
@ 2007-04-14 17:19     ` Chris Wright
  2007-04-14 17:23       ` Brian Gernhardt
  2007-04-14 19:07       ` Julian Phillips
  0 siblings, 2 replies; 16+ messages in thread
From: Chris Wright @ 2007-04-14 17:19 UTC (permalink / raw)
  To: Brian Gernhardt; +Cc: Chris Wright, Rene Herman, Linux Kernel, Greg KH, git

* Brian Gernhardt (benji@silverinsanity.com) wrote:
> On Apr 14, 2007, at 4:34 AM, Chris Wright wrote:
> >I've already put a tree like this up on kernel.org.  The master branch
> >is Linus' tree, and there's branches for each of the stable releases
> >called linux-2.6.[12-20].y (I didn't add 2.6.11.y).
> >
> >http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6-stable.git;a=summary
> 
> Is HEAD for that repo the most recent stable branch, or (as gitweb  
> makes it look) Linus's head.  I'd expect a "-stable" repo to point at  
> the most recent stable commit, not the most recent development  
> commit.  And I'd also expect gitweb's summary page to show the  
> shortlog for HEAd.  One of my assumptions are being broken and I  
> don't like it.  It leaves me all confused...

As I mentioned.  The master branch (HEAD) is Linus' tree, and each
stable tree is on its own branch.  You'll find shortlog summarizes the
main branch, so yes, gitweb's summary is a bit confusing based on your
assumptions.  This is a new tree and hasn't been publicized until now.
It does make sense to have its head be the newest stable, I'll switch
that around.

thanks,
-chris

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14 17:19     ` Chris Wright
@ 2007-04-14 17:23       ` Brian Gernhardt
  2007-04-14 17:31         ` Chris Wright
  2007-04-14 19:07       ` Julian Phillips
  1 sibling, 1 reply; 16+ messages in thread
From: Brian Gernhardt @ 2007-04-14 17:23 UTC (permalink / raw)
  To: Chris Wright; +Cc: Rene Herman, Linux Kernel, Greg KH, git


On Apr 14, 2007, at 1:19 PM, Chris Wright wrote:

> As I mentioned.  The master branch (HEAD) is Linus' tree, and each
> stable tree is on its own branch.

Hm.  I should drink my coffee, then read the mailing list, not the  
other way around.  If you do update HEAD to be the most recent  
stable, my thanks.  Apologies for the noise.

~~ Brian

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14 17:23       ` Brian Gernhardt
@ 2007-04-14 17:31         ` Chris Wright
  0 siblings, 0 replies; 16+ messages in thread
From: Chris Wright @ 2007-04-14 17:31 UTC (permalink / raw)
  To: Brian Gernhardt; +Cc: Chris Wright, Rene Herman, Linux Kernel, Greg KH, git

* Brian Gernhardt (benji@silverinsanity.com) wrote:
> Hm.  I should drink my coffee, then read the mailing list, not the  
> other way around.  If you do update HEAD to be the most recent  
> stable, my thanks.  Apologies for the noise.

No problem, was a good suggestion (and it's done).

thanks,
-chris

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14  6:24 ` Junio C Hamano
  2007-04-14  7:13   ` Rene Herman
  2007-04-14  7:21   ` Greg KH
@ 2007-04-14 17:34   ` Linus Torvalds
  2007-04-14 18:45     ` Gerb Stralko
  2 siblings, 1 reply; 16+ messages in thread
From: Linus Torvalds @ 2007-04-14 17:34 UTC (permalink / raw)
  To: Junio C Hamano; +Cc: Rene Herman, Linux Kernel, Greg KH, git



On Fri, 13 Apr 2007, Junio C Hamano wrote:
> 
> I think adding these lines to .git/config would do the trick,
> after you have done the "checkout -b v2.6.20 v2.6.20" step:
> 
> [branch "v2.6.20"]
> 	remote = stable
> 	merge = refs/heads/master

Please don't do this. Using the same name for a branch as for a tag is 
madness. Call it "v2.6.20-stable" or anything else, but don't re-use the 
same naming as for tags.

Sure, git will have some random well-defined order of parsing which one 
"v2.6.20" actually means in any particular context (usually the tag-name 
will take precedence, except for contexts where the branchname is 
required!), and yeah, you can always make things unambiguous by specifying 
the full name (ie say "[refs/]tags/v2.6.20" when you want to make sure 
it's unambiguously the tag, not the branch) but it's still guaranteed to 
cause confusion at *some* level, if only for the user.

So just don't do it. 

		Linus

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14 17:34   ` Linus Torvalds
@ 2007-04-14 18:45     ` Gerb Stralko
  0 siblings, 0 replies; 16+ messages in thread
From: Gerb Stralko @ 2007-04-14 18:45 UTC (permalink / raw)
  To: Linus Torvalds; +Cc: Junio C Hamano, Rene Herman, Linux Kernel, Greg KH, git

> Please don't do this. Using the same name for a branch as for a tag is
> madness. Call it "v2.6.20-stable" or anything else, but don't re-use the
> same naming as for tags.
>

Yes I have done this before, and it took me awhile to realize what was
going on.  It caused me some grief, and a few hours of lost time...Of
course this was back in my earlier days ;)

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14 17:19     ` Chris Wright
  2007-04-14 17:23       ` Brian Gernhardt
@ 2007-04-14 19:07       ` Julian Phillips
  1 sibling, 0 replies; 16+ messages in thread
From: Julian Phillips @ 2007-04-14 19:07 UTC (permalink / raw)
  To: Chris Wright; +Cc: Brian Gernhardt, Rene Herman, Linux Kernel, Greg KH, git

On Sat, 14 Apr 2007, Chris Wright wrote:

> * Brian Gernhardt (benji@silverinsanity.com) wrote:
>> On Apr 14, 2007, at 4:34 AM, Chris Wright wrote:
>>> I've already put a tree like this up on kernel.org.  The master branch
>>> is Linus' tree, and there's branches for each of the stable releases
>>> called linux-2.6.[12-20].y (I didn't add 2.6.11.y).
>>>
>>> http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6-stable.git;a=summary
>>
>> Is HEAD for that repo the most recent stable branch, or (as gitweb
>> makes it look) Linus's head.  I'd expect a "-stable" repo to point at
>> the most recent stable commit, not the most recent development
>> commit.  And I'd also expect gitweb's summary page to show the
>> shortlog for HEAd.  One of my assumptions are being broken and I
>> don't like it.  It leaves me all confused...
>
> As I mentioned.  The master branch (HEAD) is Linus' tree, and each
> stable tree is on its own branch.  You'll find shortlog summarizes the
> main branch, so yes, gitweb's summary is a bit confusing based on your
> assumptions.  This is a new tree and hasn't been publicized until now.
> It does make sense to have its head be the newest stable, I'll switch
> that around.

Would it not make more sense to point HEAD at the linux-2.6.20-y branch 
and either let master be Linus' tree or simply not have a master branch? 
Otherwise, what happens to master when the latest stable tree becomes 
linux-2.6.21-y?

-- 
Julian

  ---
Most people want either less corruption or more of a chance to
participate in it.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14  8:54   ` Rene Herman
@ 2007-04-14 23:13     ` Rene Herman
  2007-04-15  0:23       ` Julian Phillips
  0 siblings, 1 reply; 16+ messages in thread
From: Rene Herman @ 2007-04-14 23:13 UTC (permalink / raw)
  To: Chris Wright; +Cc: Linux Kernel, Greg KH, git

On 04/14/2007 10:54 AM, Rene Herman wrote:

> On 04/14/2007 10:34 AM, Chris Wright wrote:
> 
>> I've already put a tree like this up on kernel.org.  The master branch
>> is Linus' tree, and there's branches for each of the stable releases
>> called linux-2.6.[12-20].y (I didn't add 2.6.11.y).
>>
>> http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6-stable.git;a=summary 
> 
> I see, thank you; that sounds like a good "master" repo to clone then.

Okay, I just cloned this repo, like:

git clone -n \
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6-stable.git \
stable

How do I now checkout for example 2.6.20.6, or get a diff between 2.6.20.6 
and the (at the moment of writing latest -stable) 2.6.20.7?

rene@7ixe4:~/src/linux$ cd stable
rene@7ixe4:~/src/linux/stable$ git branch -a
* master
   origin/HEAD
   origin/linux-2.6
   origin/linux-2.6.12.y
   origin/linux-2.6.13.y
   origin/linux-2.6.14.y
   origin/linux-2.6.15.y
   origin/linux-2.6.16.y
   origin/linux-2.6.17.y
   origin/linux-2.6.18.y
   origin/linux-2.6.19.y
   origin/linux-2.6.20.y
   origin/master

and I can check them out like

rene@7ixe4:~/src/linux/stable$ git checkout origin/linux-2.6.20.y

or in this case, but only this case, like:

rene@7ixe4:~/src/linux/stable$ git checkout v2.6.20.7

"v2.6.20.7" seems to be the only tag from the stable branches that's present 
  in this tree?

rene@7ixe4:[...]$ git tag -l | grep "v2\.6\.[[:digit:]]\{1,2\}\."
v2.6.20.7

Rene.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: GIT and the current -stable
  2007-04-14 23:13     ` Rene Herman
@ 2007-04-15  0:23       ` Julian Phillips
  0 siblings, 0 replies; 16+ messages in thread
From: Julian Phillips @ 2007-04-15  0:23 UTC (permalink / raw)
  To: Rene Herman; +Cc: Chris Wright, Linux Kernel, Greg KH, git

On Sun, 15 Apr 2007, Rene Herman wrote:

>
> "v2.6.20.7" seems to be the only tag from the stable branches that's present
>  in this tree?
>
> rene@7ixe4:[...]$ git tag -l | grep "v2\.6\.[[:digit:]]\{1,2\}\."
> v2.6.20.7

Obviously I don't know how Chris created his conglomerated repo, but I 
just made one of my own, and it has all the tags I would expect in it ... 
so it's not an inherent git problem (or not in 1.5.1.1 anyway).

I guess that Chris created his in such a way that the automated tag 
following code didn't trigger? (Or maybe used a really old git?)

(Mine's at http://git.q42.co.uk/w/stable.git if anyone is interested 
enough to want to look at it ...)

It only took me 4 commands to create too (ok, so three of them were bash 
for loops ... and I did do a little bit more to tidy up), I do enjoy using 
flexible tools :D.

-- 
Julian

  ---
Go slowly to the entertainments of thy friends, but quickly to their
misfortunes.
 		-- Chilo

^ permalink raw reply	[flat|nested] 16+ messages in thread

end of thread, other threads:[~2007-04-15  0:24 UTC | newest]

Thread overview: 16+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-04-14  5:36 GIT and the current -stable Rene Herman
2007-04-14  6:24 ` Junio C Hamano
2007-04-14  7:13   ` Rene Herman
2007-04-14  7:21   ` Greg KH
2007-04-14 17:34   ` Linus Torvalds
2007-04-14 18:45     ` Gerb Stralko
2007-04-14  8:34 ` Chris Wright
2007-04-14  8:54   ` Rene Herman
2007-04-14 23:13     ` Rene Herman
2007-04-15  0:23       ` Julian Phillips
2007-04-14 15:15   ` Brian Gernhardt
2007-04-14 17:19     ` Chris Wright
2007-04-14 17:23       ` Brian Gernhardt
2007-04-14 17:31         ` Chris Wright
2007-04-14 19:07       ` Julian Phillips
2007-04-14 11:56 ` Uwe Kleine-König

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).