Re: Stable -rc git trees and email headers

[Kevin Hilman]

On Wed, Jun 22, 2016 at 8:14 AM, Willy Tarreau <w@xxxxxx> wrote:
> On Wed, Jun 22, 2016 at 08:11:41AM -0700, Greg KH wrote:
>> On Wed, Jun 22, 2016 at 08:02:14AM -0700, Kevin Hilman wrote:
>> > Hi Greg,
>> >
>> > On Fri, Jun 17, 2016 at 8:16 PM, Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
>> >
>> > [...]
>> >
>> > > Kevin, hopefully this tree also helps you.  I will also be adding some
>> > > email headers to my announcements of the -rc releases that you can
>> > > parse.  Here is what they are going to look like for an example release:
>> > >
>> > > X-KernelTest-Patch: http://kernel.org/pub/linux/kernel/v4.x/stable-review/patch-4.4.14-rc1.gz
>> > > X-KernelTest-Tree: git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git
>> > > X-KernelTest-Branch: linux-4.4.y
>> > > X-KernelTest-Patches: git://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git
>> > > X-KernelTest-Version: 4.4.14-rc1
>> > > X-KernelTest-Deadline: 2016-06-20T02:57+00:00
>> > >
>> > > Will those work out?
>> >
>> > Can you add one more:  X-KernelTest-PatchCount ?
>> >
>> > Which would include the total number of patches in the series/branch?
>> > This will help us determine uniqueness for a given branch.
>>
>> What do you mean by "uniqueness"?
>>
>> As these are going to be the -rc releases, they will be unique from any
>> other -rc release.
>
> I guess he means between incremental updates, though that doesn't work
> if you drop a patch.

Right, you add/drop patches as various regressions come in, so the
branch is going to keep changing, even though it's going to continue
to be called -rc1.

Actually, now that you're using git trees, even better than patch
count would be just to publish the result of 'git describe' on the
branch you've pushed.  That includes the patch count and the has of
the top commit.

X-KernelTest-GitDescribe?

Kevin