Wolfgang Grandegger wrote:
> Sebastian Haas wrote:
>> Wolfgang Grandegger schrieb:
>>> Sebastian Haas wrote:
>>>> Marc Kleine-Budde schrieb:
>>>>>>>> "static struct sk_buff *". Patches are welcome.
>>>>>>> fixed. patches against which tree?
>>>>>> First again SVN trunk.
>>>>> What about a git tree? The whole svn situation suxs.
>>> What should be in the git repository? Wht do you expect? Maintaining a
>>> repository is real work, which I would like to minimize. Volunteers are
>>> always welcome, of course.
>> Here I am. Would be glad to support Socket CAN in this way.
> 
> I will keep in mind.
> 
>>>> Full ack. Submitting a patch with git is just a matter of seconds while
>>>> with svn it takes minutes to copy and paste everything together in
>>>> Thunderbird.
>>> That's not a real pro for using a git ;-). Will comment in more detail
>>> tomorrow.
>> Yes, I know. For externals with write permission to git is a great 
>> pleasure. You can work on a feature, check in your changes and simply 
>> create patches.
> 
> You can checkout Dave's net-next-2.6 branch. What are you missing? I see
> three options here:
> 
> 1. Setup can-next-2.6 to ease the maintenance of David Miller. I would
>    apply *accepted* patches to this branch and tell Dave from time to
>    time to pull from it. There is little benefit for normal users.

And also for Dave as he stated here:

http://marc.info/?l=linux-netdev&m=125207489710387&w=2

> 
> 2. Setup a full can-2.6 git tree with *all* drivers from the SVN
>    repository (excluding the old ones). That's a significant maintenance
>    effort, especially to provide compatibility with older versions.

IMO it does not make any sense to maintain backward compatible code in a brand
new kernel tree environment ...

> 
> 3. Move the out-of-tree code from the BerliOS SVN repository to git.
>    That's a valuable option, indeed.

Ack.

For me the current process could be:

* For new drivers (like ems_usb.c):

1. Post source as a diff to net-next-2.6 on SocketCAN Core ML
2. Make some reviews. Still needed changes: goto 1
3. Post source as a diff to net-next-2.6 on netdev ML
4. Make some reviews. Still needed changes: goto 1 or 3
5. Acked-by: Wolfgang
6. Done Mainlining
7. Apply source to out-of-tree-repository
8. Add backward compatibility stuff
9. Done


* For drivers that are only in the out-of-tree-repository:

1. Post patches as a diff to out-of-tree-repository on SocketCAN Core ML
2. Make some reviews. Still needed changes: goto 1
3. Apply to the out-of-tree-repository
4. Done


* For drivers that are in Mainline (and in the out-of-tree-repository):

Post the patches/fixes for the Mainline driver as described in the process for
new drivers.


* To mainline drivers that are currently only in the out-of-tree-repository:

Create a source from the out-of-tree-repository with strip-src and follow the
process for new drivers.


This should make it.

And in most cases the developer can post git-patches based on net-next-2.6
which becomes the leading development repository.

It is no big difference then whether the out-of-tree-repository is a SVN or a
GIT then. I could live with both of them.

Looks the described process sane to all of you?

Regards,
Oliver

_______________________________________________
Socketcan-core mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/socketcan-core

Reply via email to