2004-09-14 21:36:31

by Daniel Andersen

[permalink] [raw]
Subject: [PATCH] README (resend) - Explain new 2.6.xx.x version number

This one ended up in the void last time without any comments.

Daniel Andersen

--

diff -urN linux/README.orig linux/README
--- linux/README.orig 2004-08-14 07:37:40.000000000 +0200
+++ linux/README 2004-08-27 17:16:55.107413637 +0200
@@ -76,6 +76,23 @@
the backup files (xxx~ or xxx.orig), and make sure that there are no
failed patches (xxx# or xxx.rej). If there are, either you or me has
made a mistake.
+
+ As of kernel 2.6.8 there was a bug-fix release numbering scheme
+ introduced. In such cases a fourth number is added to the release
+ version, eg. 2.6.8.1. When patching from a 2.6.xx(.x) release to a
+ newer version, patches are to be applied against the original
+ release, eg. 2.6.8 and not the bug-fix release 2.6.8.1. In case of a
+ bug-fix release such as if eg. 2.6.8.2 is released after 2.6.9 has
+ been released, 2.6.9 is still to be considered the newest kernel
+ release of all current kernels. Old patches can be reversed by
+ adding the "-R" option to the patch tool.
+
+ Example to apply a bugfix release patch:
+ bzip2 -dc ../patch-2.6.8.1.bz2 | patch -p1
+
+ Example to apply a new release on a bugfix tree:
+ bzip2 -dc ../patch-2.6.8.1.bz2 | patch -p1 -R
+ bzip2 -dc ../patch-2.6.9.bz2 | patch -p1

Alternatively, the script patch-kernel can be used to automate this
process. It determines the current kernel version and applies any


2004-09-14 21:58:45

by Dave Jones

[permalink] [raw]
Subject: Re: [PATCH] README (resend) - Explain new 2.6.xx.x version number

On Tue, Sep 14, 2004 at 11:35:15PM +0200, Daniel Andersen wrote:

> + In case of a
> + bug-fix release such as if eg. 2.6.8.2 is released after 2.6.9 has
> + been released, 2.6.9 is still to be considered the newest kernel
> + release of all current kernels.

This bit seems odd to me. Why would a 2.6.8.2 get released, when there's
a newer 2.6.9 which should fix whatever was relevant to get into 2.6.8.x ?

Dave

2004-09-14 22:29:36

by Daniel Andersen

[permalink] [raw]
Subject: Re: [PATCH] README (resend) - Explain new 2.6.xx.x version number

Dave Jones wrote:
> On Tue, Sep 14, 2004 at 11:35:15PM +0200, Daniel Andersen wrote:
>
> > + In case of a
> > + bug-fix release such as if eg. 2.6.8.2 is released after 2.6.9 has
> > + been released, 2.6.9 is still to be considered the newest kernel
> > + release of all current kernels.
>
> This bit seems odd to me. Why would a 2.6.8.2 get released, when there's
> a newer 2.6.9 which should fix whatever was relevant to get into 2.6.8.x ?
>
> Dave
>

This was discussed in the thread "Linux 2.6.9-rc1". Linus said there was
a remote possibility it could happen some day so I thought it was a good
thing to mention.

Daniel Andersen

--

2004-09-15 01:03:51

by Roman Zippel

[permalink] [raw]
Subject: Re: [PATCH] README (resend) - Explain new 2.6.xx.x version number

Hi,

On Tue, 14 Sep 2004, Daniel Andersen wrote:

> This one ended up in the void last time without any comments.

The funny thing is by the time people managed to apply the patch
correctly, they don't need to read the README anymore.
Seriously, without knowing about the pre-patches, what would you expect
about the patch order if you found the patches 2.6.8, 2.6.8.1, 2.6.9?

bye, Roman

2004-09-15 02:23:40

by Daniel Andersen

[permalink] [raw]
Subject: Re: [PATCH] README (resend) - Explain new 2.6.xx.x version number

Roman Zippel wrote:
> Hi,
>
> On Tue, 14 Sep 2004, Daniel Andersen wrote:
>
>
>>This one ended up in the void last time without any comments.
>
>
> The funny thing is by the time people managed to apply the patch
> correctly, they don't need to read the README anymore.
> Seriously, without knowing about the pre-patches, what would you expect
> about the patch order if you found the patches 2.6.8, 2.6.8.1, 2.6.9?
>
> bye, Roman

I'm not sure what you really mean to say. And I'm not expecting
anything. I wrote the patch because I see people ask, even on this list,
if 2.6.9 is to be applied against 2.6.8 or 2.6.8.1. Whats the trouble?

Daniel Andersen

--

2004-09-15 02:42:11

by Randy.Dunlap

[permalink] [raw]
Subject: Re: [PATCH] README (resend) - Explain new 2.6.xx.x version number

> Hi,
>
> On Tue, 14 Sep 2004, Daniel Andersen wrote:
>
>> This one ended up in the void last time without any comments.
>
> The funny thing is by the time people managed to apply the patch
> correctly, they don't need to read the README anymore.

That's correct for this time. However, if they have other kernel
trees (in the future) with this patch applied, it can help.

> Seriously, without knowing about the pre-patches, what would you expect
> about the patch order if you found the patches 2.6.8, 2.6.8.1, 2.6.9?

We have evidence that it's confusing to more than one person.

~Randy

2004-09-15 08:01:59

by Denis Vlasenko

[permalink] [raw]
Subject: Re: [PATCH] README (resend) - Explain new 2.6.xx.x version number

On Wednesday 15 September 2004 05:40, Randy.Dunlap wrote:
> > Hi,
> >
> > On Tue, 14 Sep 2004, Daniel Andersen wrote:
> >> This one ended up in the void last time without any comments.
> >
> > The funny thing is by the time people managed to apply the patch
> > correctly, they don't need to read the README anymore.
>
> That's correct for this time. However, if they have other kernel
> trees (in the future) with this patch applied, it can help.
>
> > Seriously, without knowing about the pre-patches, what would you expect
> > about the patch order if you found the patches 2.6.8, 2.6.8.1, 2.6.9?
>
> We have evidence that it's confusing to more than one person.

Using 2.6.8-fix1 instead of 2.6.8.1 could avoid such problems.
--
vda

2004-09-15 11:06:38

by Roman Zippel

[permalink] [raw]
Subject: Re: [PATCH] README (resend) - Explain new 2.6.xx.x version number

Hi,

On Tue, 14 Sep 2004, Randy.Dunlap wrote:

> > Seriously, without knowing about the pre-patches, what would you expect
> > about the patch order if you found the patches 2.6.8, 2.6.8.1, 2.6.9?
>
> We have evidence that it's confusing to more than one person.

The question is what is more confusing to most of the users?
What are the expectation of the average user?

bye, Roman

2004-09-15 15:29:38

by Randy.Dunlap

[permalink] [raw]
Subject: Re: [PATCH] README (resend) - Explain new 2.6.xx.x version number

On Wed, 15 Sep 2004 13:06:06 +0200 (CEST) Roman Zippel wrote:

| Hi,
|
| On Tue, 14 Sep 2004, Randy.Dunlap wrote:
|
| > > Seriously, without knowing about the pre-patches, what would you expect
| > > about the patch order if you found the patches 2.6.8, 2.6.8.1, 2.6.9?
| >
| > We have evidence that it's confusing to more than one person.
|
| The question is what is more confusing to most of the users?
| What are the expectation of the average user?

I have no idea. A decision was made and now there are some
attempts to document it. Are you trying to ask for the decision
to be reconsidered?

--
~Randy

2004-09-15 16:00:23

by Roman Zippel

[permalink] [raw]
Subject: Re: [PATCH] README (resend) - Explain new 2.6.xx.x version number

Hi,

On Wed, 15 Sep 2004, Randy.Dunlap wrote:

> I have no idea. A decision was made and now there are some
> attempts to document it.

When and by whom?

> Are you trying to ask for the decision
> to be reconsidered?

I'd really be interested in the reasoning for it and against the
alternatives.

bye, Roman