[ANNOUNCE] Switching to qemu-xen.git and qemu-xen-traditional.git

[ANNOUNCE] Switching to qemu-xen.git and qemu-xen-traditional.git

[Ian Campbell]

I have just completed the switch to use a single qemu tree for each of the
versions we support (qemu-xen AKA upstream and qemu-xen-traditional AKA our
historical fork).

Instead of the individual Xen release specific repositories the QEMU trees
are now:

http://xenbits.xen.org/gitweb/?p=qemu-xen-traditional.git
http://xenbits.xen.org/gitweb/?p=qemu-xen.git

(see those pages for the new URLs). Each Xen release now corresponds to
branches within those repositories.

For qemu-xen.git there are staging+master and staging-X.Y+stable-X.Y
branches, representing tested and untested (by osstest push gate).

For qemu-xen-traditional.git there are only master and stable-X.Y branches.

The changes to use these trees instead of the old ones are now in
xen.git#staging and will be backported to stable branches and releases in
due course.

The old qemu-upstream stable qemu-upstream-X.Y-testing.git repositories are
being kept for the time being and are updated by osstest in lockstep with
the branches in the new tree via the relevant flights.

The old qemu-traditional stable repositories qemu-xen-X.Y-testing.git are
being kept for the time being and Ian J will push to them along with the
new branches (these trees do not have a push gate).

All of these will be removed eventually, i.e. at the soonest when a
relevant stable release has occurred.

The old development repos (qemu-xen=qemu-upstream-unstable and qemu-xen
-traditional=qemu-xen-unstable.git) remain for now but will not see any
further updates. These will be removed soon, once osstest no longer
references them for bisects.

All of the staging/qemu-* trees have already been removed.

Ian

Killed osstest flights (Was: Re: [ANNOUNCE] Switching to qemu-xen.git and qemu-xen-traditional.git)

[Ian Campbell]

On Wed, 2015-10-21 at 14:16 +0100, Ian Campbell wrote:
> I have just completed the switch to use a single qemu tree for each of the
> versions we support (qemu-xen AKA upstream and qemu-xen-traditional AKA our
> historical fork).

Some flights which were active during the transition have been killed since
their build phase failed. I should have stopped everything and not just the
qemu flights, sorry.

The affected flights are 63102 63107 63113 and 63115 from the main colo
instance.

Also 38187, 38189 and 38190 from the Cambridge instance.

Sorry for the noise.

Ian.

Re: [ANNOUNCE] Switching to qemu-xen.git and qemu-xen-traditional.git

[Olaf Hering]

On Wed, Oct 21, Ian Campbell wrote:

> http://xenbits.xen.org/gitweb/?p=qemu-xen.git

Is the HEAD warning expected?

$ git clone git://xenbits.xen.org/qemu-xen qemu-xen.git
$ cd $_
$ git remote show origin
* remote origin
  Fetch URL: git://xenbits.xen.org/qemu-xen
  Push  URL: git://xenbits.xen.org/qemu-xen
  HEAD branch (remote HEAD is ambiguous, may be one of the following):
    master
    staging
  Remote branches:
    master          tracked
    stable-4.2      tracked
    stable-4.3      tracked
    stable-4.4      tracked
    stable-4.5      tracked
    stable-4.6      tracked
    staging         tracked
    staging-4.2     tracked
    staging-4.3     tracked
    staging-4.4     tracked
    staging-4.5     tracked
    staging-4.6     tracked
    upstream-tested tracked
  Local branch configured for 'git pull':
    master merges with remote master
  Local ref configured for 'git push':
    master pushes to master (up to date)

Olaf

Re: [ANNOUNCE] Switching to qemu-xen.git and qemu-xen-traditional.git

[Ian Campbell]

On Wed, 2015-10-21 at 17:26 +0200, Olaf Hering wrote:
> On Wed, Oct 21, Ian Campbell wrote:
> 
> > http://xenbits.xen.org/gitweb/?p=qemu-xen.git
> 
> Is the HEAD warning expected?

No it is not.

Both trees contain a file HEAD containing:
    ref: refs/heads/master
which is the same as xen.git does.

I see the same thing as you for the qemu-xen tree, but not for xen.git. I'm
not sure what I'm supposed to do about this.

Strangely I don't see this on my http based clone of qemu-xen.git (this is
not a solution though!).

I shall consult around and see if I can figure out what I'm supposed to do
here.

Ian.

Re: [ANNOUNCE] Switching to qemu-xen.git and qemu-xen-traditional.git

[Ian Campbell]

On Wed, 2015-10-21 at 16:52 +0100, Ian Campbell wrote:
> On Wed, 2015-10-21 at 17:26 +0200, Olaf Hering wrote:
> > On Wed, Oct 21, Ian Campbell wrote:
> > 
> > > http://xenbits.xen.org/gitweb/?p=qemu-xen.git
> > 
> > Is the HEAD warning expected?
> 
> No it is not.

Apparently it actually is, it's a short coming of the git:// protocol that
the client only sees the sha which HEAD points to and has to reverse
engineer which branch it was, which is ambiguous if there are two such
branches (as there is in this repo right now).

Ian J found this which explains things a bit better:

http://stackoverflow.com/questions/13061364/why-git-remote-show-name-display-remote-head-is-ambiguous-may-be-one-of-the

> 
> Both trees contain a file HEAD containing:
>     ref: refs/heads/master
> which is the same as xen.git does.

> I see the same thing as you for the qemu-xen tree, but not for xen.git.


In xen.git staging and master happen to differ right now, so we don't see
this there.

> I shall consult around and see if I can figure out what I'm supposed to do
> here.

a) Nothing, sorry.

Maybe one day a) will be "update git on xenbits", I don't intend to even
investigate that right now, for hopefully obvious reasons...

Ian.