linux-2.6.21-gitN - versioning question

linux-2.6.21-gitN - versioning question

[Jan Engelhardt]

Hi list,


I notice that people refer to certain git snapshots as e.g. -git16; 
kernel.org does so too. Hovering over the link on kernel.org reveals 
http://kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.21-git16.bz2
but where do I actually find -git16 in Linus's git tree? git16 is 
neither a branch nor a tag.


	Jan
-- 

Re: linux-2.6.21-gitN - versioning question

[Adrian Bunk]

On Sat, May 12, 2007 at 01:19:23PM +0200, Jan Engelhardt wrote:
> Hi list,
> 
> 
> I notice that people refer to certain git snapshots as e.g. -git16; 
> kernel.org does so too. Hovering over the link on kernel.org reveals 
> http://kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.21-git16.bz2
> but where do I actually find -git16 in Linus's git tree? git16 is 
> neither a branch nor a tag.

http://kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.21-git16.id

> 	Jan

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: linux-2.6.21-gitN - versioning question

[YOSHIFUJI Hideaki / 吉藤英明]

In article <Pine.LNX.4.61.0705121316540.9570@yvahk01.tjqt.qr> (at Sat, 12 May 2007 13:19:23 +0200 (MEST)), Jan Engelhardt <jengelh@linux01.gwdg.de> says:

> I notice that people refer to certain git snapshots as e.g. -git16; 
> kernel.org does so too. Hovering over the link on kernel.org reveals 
> http://kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.21-git16.bz2
> but where do I actually find -git16 in Linus's git tree? git16 is 
> neither a branch nor a tag.

http://kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.21-git16.id
tells the commit in his tree.

--yoshfuji

Re: linux-2.6.21-gitN - versioning question

[Cyrill Gorcunov]

[YOSHIFUJI Hideaki / ?$B5HF#1QL@ - Sat, May 12, 2007 at 09:03:08PM +0900]
| In article <Pine.LNX.4.61.0705121316540.9570@yvahk01.tjqt.qr> (at Sat, 12 May 2007 13:19:23 +0200 (MEST)), Jan Engelhardt <jengelh@linux01.gwdg.de> says:
| 
| > I notice that people refer to certain git snapshots as e.g. -git16; 
| > kernel.org does so too. Hovering over the link on kernel.org reveals 
| > http://kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.21-git16.bz2
| > but where do I actually find -git16 in Linus's git tree? git16 is 
| > neither a branch nor a tag.
| 
| http://kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.21-git16.id
| tells the commit in his tree.
| 
| --yoshfuji

Actually I think it would be convenient if such tags (like v.2.6.21-git16)
were in Linus' git tree too.


		Cyrill

Re: linux-2.6.21-gitN - versioning question

[David Woodhouse]

On Sat, 2007-05-12 at 17:19 +0400, Cyrill Gorcunov wrote:
> Actually I think it would be convenient if such tags (like
> v.2.6.21-git16) were in Linus' git tree too.

Then there would be _lots_ of tags in the master tree -- I'm not sure we
want that.

I suppose I could put a tree on kernel.org which _does_ have the tags,
and which precisely matches Linus' tree in every other way... but then
again, you could do that too by just rsyncing the tags yourself :)

-- 
dwmw2

Re: linux-2.6.21-gitN - versioning question

[Cyrill Gorcunov]

[David Woodhouse - Sat, May 12, 2007 at 09:44:56PM +0800]
| On Sat, 2007-05-12 at 17:19 +0400, Cyrill Gorcunov wrote:
| > Actually I think it would be convenient if such tags (like
| > v.2.6.21-git16) were in Linus' git tree too.
| 
| Then there would be _lots_ of tags in the master tree -- I'm not sure we
| want that.
| 
| I suppose I could put a tree on kernel.org which _does_ have the tags,
| and which precisely matches Linus' tree in every other way... but then
| again, you could do that too by just rsyncing the tags yourself :)
| 
| -- 
| dwmw2
| 

Hi David,

thanks for reply. Indeed ;) Leave all as it is.

		Cyrill

Re: linux-2.6.21-gitN - versioning question

[Jan Engelhardt]

On May 12 2007 21:44, David Woodhouse wrote:
>On Sat, 2007-05-12 at 17:19 +0400, Cyrill Gorcunov wrote:
>> Actually I think it would be convenient if such tags (like
>> v.2.6.21-git16) were in Linus' git tree too.
>
>Then there would be _lots_ of tags in the master tree -- I'm not sure we
>want that.
>
>I suppose I could put a tree on kernel.org which _does_ have the tags,
>and which precisely matches Linus' tree in every other way... but then
>again, you could do that too by just rsyncing the tags yourself :)

Are you saying GIT tags are expensive, compared to, say, SVN tags?


	Jan
-- 

Re: linux-2.6.21-gitN - versioning question

[Cyrill Gorcunov]

[Jan Engelhardt - Sat, May 12, 2007 at 05:01:19PM +0200]
| 
| On May 12 2007 21:44, David Woodhouse wrote:
| >On Sat, 2007-05-12 at 17:19 +0400, Cyrill Gorcunov wrote:
| >> Actually I think it would be convenient if such tags (like
| >> v.2.6.21-git16) were in Linus' git tree too.
| >
| >Then there would be _lots_ of tags in the master tree -- I'm not sure we
			^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
| >want that.
| >
| >I suppose I could put a tree on kernel.org which _does_ have the tags,
| >and which precisely matches Linus' tree in every other way... but then
| >again, you could do that too by just rsyncing the tags yourself :)
| 
| Are you saying GIT tags are expensive, compared to, say, SVN tags?
| 
| 
| 	Jan
| -- 
| 

Hi Jan,

you know the current Linus' git tree does contain ~62 tags
(if I'm right). And of coure that's very small number of files
compared to the number of blobs in the master tree. So I think the
reason to don't use a lot of tags - is just... well... a lot of
tags in the git tree ;)

[Don't beat me ;]

		Cyrill

Re: linux-2.6.21-gitN - versioning question

[Cyrill Gorcunov]

[Jan Engelhardt - Sat, May 12, 2007 at 05:01:19PM +0200]
| 
| On May 12 2007 21:44, David Woodhouse wrote:
| >On Sat, 2007-05-12 at 17:19 +0400, Cyrill Gorcunov wrote:
| >> Actually I think it would be convenient if such tags (like
| >> v.2.6.21-git16) were in Linus' git tree too.
| >
| >Then there would be _lots_ of tags in the master tree -- I'm not sure we
| >want that.
| >
| >I suppose I could put a tree on kernel.org which _does_ have the tags,
| >and which precisely matches Linus' tree in every other way... but then
| >again, you could do that too by just rsyncing the tags yourself :)
| 
| Are you saying GIT tags are expensive, compared to, say, SVN tags?
| 

Anyway, Davind wrote that he could put a special tree with tags on
kernel.org. And we always have opportunity to make a such tags
in our own tree copies by hands. But I prefer keep (on my computer)
pure Linus' git tree clone ;)

		Cyrill

Re: linux-2.6.21-gitN - versioning question

[Russell King]

On Sat, May 12, 2007 at 08:24:28PM +0400, Cyrill Gorcunov wrote:
> [Jan Engelhardt - Sat, May 12, 2007 at 05:01:19PM +0200]
> | 
> | On May 12 2007 21:44, David Woodhouse wrote:
> | >On Sat, 2007-05-12 at 17:19 +0400, Cyrill Gorcunov wrote:
> | >> Actually I think it would be convenient if such tags (like
> | >> v.2.6.21-git16) were in Linus' git tree too.
> | >
> | >Then there would be _lots_ of tags in the master tree -- I'm not sure we
> | >want that.
> | >
> | >I suppose I could put a tree on kernel.org which _does_ have the tags,
> | >and which precisely matches Linus' tree in every other way... but then
> | >again, you could do that too by just rsyncing the tags yourself :)
> | 
> | Are you saying GIT tags are expensive, compared to, say, SVN tags?

(Sorry, don't have Jan's message.)

The real issue is not that, but that David doesn't have write access to
Linus' published tree.  Neither should he because he could contaminate
Linus' published tree with the MTD tree, or introduce some unreviewed
patch which gave him backdoor access via the kernel.

(Of course, David's unlikely to do either, but that's not really the
point.)

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:

Re: linux-2.6.21-gitN - versioning question

[Cyrill Gorcunov]

[Russell King - Sat, May 12, 2007 at 05:31:41PM +0100]
| On Sat, May 12, 2007 at 08:24:28PM +0400, Cyrill Gorcunov wrote:
| > [Jan Engelhardt - Sat, May 12, 2007 at 05:01:19PM +0200]
| > | 
| > | On May 12 2007 21:44, David Woodhouse wrote:
| > | >On Sat, 2007-05-12 at 17:19 +0400, Cyrill Gorcunov wrote:
| > | >> Actually I think it would be convenient if such tags (like
| > | >> v.2.6.21-git16) were in Linus' git tree too.
| > | >
| > | >Then there would be _lots_ of tags in the master tree -- I'm not sure we
| > | >want that.
| > | >
| > | >I suppose I could put a tree on kernel.org which _does_ have the tags,
| > | >and which precisely matches Linus' tree in every other way... but then
| > | >again, you could do that too by just rsyncing the tags yourself :)
| > | 
| > | Are you saying GIT tags are expensive, compared to, say, SVN tags?
| 
| (Sorry, don't have Jan's message.)
| 
| The real issue is not that, but that David doesn't have write access to
| Linus' published tree.  Neither should he because he could contaminate
| Linus' published tree with the MTD tree, or introduce some unreviewed
| patch which gave him backdoor access via the kernel.
| 
| (Of course, David's unlikely to do either, but that's not really the
| point.)
| 
| -- 
| Russell King
|  Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
|  maintainer of:
| 


The only reason I wrote some messages to this theme is that on
my point of view the using of tags like "v2.6.21-git16" would be
convenient. But that should be Linus' decision (thanks to David
for proposed solution). Anyway if Linus doesn't want to keep
a such tags in hisown tree I prefer to do the same ;)


		Cyrill

Re: linux-2.6.21-gitN - versioning question

[H. Peter Anvin]

David Woodhouse wrote:
> On Sat, 2007-05-12 at 17:19 +0400, Cyrill Gorcunov wrote:
>> Actually I think it would be convenient if such tags (like
>> v.2.6.21-git16) were in Linus' git tree too.
> 
> Then there would be _lots_ of tags in the master tree -- I'm not sure we
> want that.
> 
> I suppose I could put a tree on kernel.org which _does_ have the tags,
> and which precisely matches Linus' tree in every other way... but then
> again, you could do that too by just rsyncing the tags yourself :)

I think that would be a great idea.  It would hardly take up any space
with proper use of alternates.

	-hpa

Re: linux-2.6.21-gitN - versioning question

[David Woodhouse]

On Sat, 2007-05-12 at 17:01 +0200, Jan Engelhardt wrote:
> On May 12 2007 21:44, David Woodhouse wrote:
> >On Sat, 2007-05-12 at 17:19 +0400, Cyrill Gorcunov wrote:
> >> Actually I think it would be convenient if such tags (like
> >> v.2.6.21-git16) were in Linus' git tree too.
> >
> >Then there would be _lots_ of tags in the master tree -- I'm not sure we
> >want that.
> >
> >I suppose I could put a tree on kernel.org which _does_ have the tags,
> >and which precisely matches Linus' tree in every other way... but then
> >again, you could do that too by just rsyncing the tags yourself :)
> 
> Are you saying GIT tags are expensive, compared to, say, SVN tags?

I wasn't saying that -- I've no idea how expensive SVN tags are. But I
was of the opinion that GIT tags aren't free -- I thought I'd seen
git-pull explicitly checking (with one round-trip latency) that it has
the corresponding commit for every tag it knows about.

-- 
dwmw2

Re: linux-2.6.21-gitN - versioning question

[Cyrill Gorcunov]

[David Woodhouse - Sun, May 13, 2007 at 02:26:40PM +0800]
| On Sat, 2007-05-12 at 17:01 +0200, Jan Engelhardt wrote:
| > On May 12 2007 21:44, David Woodhouse wrote:
| > >On Sat, 2007-05-12 at 17:19 +0400, Cyrill Gorcunov wrote:
| > >> Actually I think it would be convenient if such tags (like
| > >> v.2.6.21-git16) were in Linus' git tree too.
| > >
| > >Then there would be _lots_ of tags in the master tree -- I'm not sure we
| > >want that.
| > >
| > >I suppose I could put a tree on kernel.org which _does_ have the tags,
| > >and which precisely matches Linus' tree in every other way... but then
| > >again, you could do that too by just rsyncing the tags yourself :)
| > 
| > Are you saying GIT tags are expensive, compared to, say, SVN tags?
| 
| I wasn't saying that -- I've no idea how expensive SVN tags are. But I
| was of the opinion that GIT tags aren't free -- I thought I'd seen
| git-pull explicitly checking (with one round-trip latency) that it has
| the corresponding commit for every tag it knows about.
| 
| -- 
| dwmw2
| 

Btw David who does release a such snapshots (like git16)? You or Linus?

		Cyrill

Re: linux-2.6.21-gitN - versioning question

[David Woodhouse]

On Sun, 2007-05-13 at 10:43 +0400, Cyrill Gorcunov wrote:
> Btw David who does release a such snapshots (like git16)? You or
> Linus? 

I do. Not that I've actually had much involvement with it for a while --
it's mostly done by cron and the script at
http://david.woodhou.se/git-snapshot.sh

-- 
dwmw2

Re: linux-2.6.21-gitN - versioning question

[Cyrill Gorcunov]

[David Woodhouse - Sun, May 13, 2007 at 02:50:10PM +0800]
| On Sun, 2007-05-13 at 10:43 +0400, Cyrill Gorcunov wrote:
| > Btw David who does release a such snapshots (like git16)? You or
| > Linus? 
| 
| I do. Not that I've actually had much involvement with it for a while --
| it's mostly done by cron and the script at
| http://david.woodhou.se/git-snapshot.sh
| 
| -- 
| dwmw2
| 

Understood. Thanks.

		Cyrill