All of lore.kernel.org
 help / color / mirror / Atom feed
* New developing branch of grub2
@ 2009-07-17  9:52 Bean
  2009-07-17 14:24 ` Pavel Roskin
  0 siblings, 1 reply; 3+ messages in thread
From: Bean @ 2009-07-17  9:52 UTC (permalink / raw)
  To: The development of GRUB 2

Hi,

I've created a repository at GitHub to hold some developing patches,
the main repos is at:

http://github.com/grub/grub/

master is the developing branch, while svn is the mirror of grub2 svn.

I also have a forked project at:

http://github.com/bean123/grub/

The lib branch contains the new object format code.

-- 
Bean



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: New developing branch of grub2
  2009-07-17  9:52 New developing branch of grub2 Bean
@ 2009-07-17 14:24 ` Pavel Roskin
  2009-07-17 16:24   ` Bean
  0 siblings, 1 reply; 3+ messages in thread
From: Pavel Roskin @ 2009-07-17 14:24 UTC (permalink / raw)
  To: The development of GRUB 2

On Fri, 2009-07-17 at 17:52 +0800, Bean wrote:
> Hi,
> 
> I've created a repository at GitHub to hold some developing patches,
> the main repos is at:
> 
> http://github.com/grub/grub/
> 
> master is the developing branch, while svn is the mirror of grub2 svn.
> 
> I also have a forked project at:
> 
> http://github.com/bean123/grub/
> 
> The lib branch contains the new object format code.

It's hard to see the unmerged changes in those repositories.

I can tell from my experience with Linux kernel development that
branches are not used for such efforts.  Patches need to be posted,
reviewed and discussed by others.

Separate branches or repositories are used for subsystems that have
their own maintainers.  That is, there is a repository for wireless
networking.  Patches are still posted and reviewed in a separate list.
Once they are good, they are committed to the subsystem repository.  And
then Linus pulls from it because he trusts the subsystem maintainers and
knows that the patches have been reviewed.

When patches are publishing for review, there is incentive for the patch
author to make changes clear.  When the patches are published as a
branch, chances are that the development will go too far before other
developers have a look at the code.

-- 
Regards,
Pavel Roskin



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: New developing branch of grub2
  2009-07-17 14:24 ` Pavel Roskin
@ 2009-07-17 16:24   ` Bean
  0 siblings, 0 replies; 3+ messages in thread
From: Bean @ 2009-07-17 16:24 UTC (permalink / raw)
  To: The development of GRUB 2

On Fri, Jul 17, 2009 at 10:24 PM, Pavel Roskin<proski@gnu.org> wrote:
> On Fri, 2009-07-17 at 17:52 +0800, Bean wrote:
>> Hi,
>>
>> I've created a repository at GitHub to hold some developing patches,
>> the main repos is at:
>>
>> http://github.com/grub/grub/
>>
>> master is the developing branch, while svn is the mirror of grub2 svn.
>>
>> I also have a forked project at:
>>
>> http://github.com/bean123/grub/
>>
>> The lib branch contains the new object format code.
>
> It's hard to see the unmerged changes in those repositories.
>
> I can tell from my experience with Linux kernel development that
> branches are not used for such efforts.  Patches need to be posted,
> reviewed and discussed by others.
>
> Separate branches or repositories are used for subsystems that have
> their own maintainers.  That is, there is a repository for wireless
> networking.  Patches are still posted and reviewed in a separate list.
> Once they are good, they are committed to the subsystem repository.  And
> then Linus pulls from it because he trusts the subsystem maintainers and
> knows that the patches have been reviewed.
>
> When patches are publishing for review, there is incentive for the patch
> author to make changes clear.  When the patches are published as a
> branch, chances are that the development will go too far before other
> developers have a look at the code.

Hi,

Thanks a lot for the advice ! The above development modal seems fine,
but github also provides similar technique. Users can work on their
own fork, and then send pull request to other developers whom might be
interested. Then a maintainer can merge them into the main repo. As
grub is relatively small, I think this should be enough.

The subsystem is also nice, but the prerequisite is that grub can
split into different components and compile separately. Right now, the
conf file are central and modules must be compiled at the same time,
this is not good for subsystem.

-- 
Bean



^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2009-07-17 16:24 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2009-07-17  9:52 New developing branch of grub2 Bean
2009-07-17 14:24 ` Pavel Roskin
2009-07-17 16:24   ` Bean

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.