Branching

Branching

[Chong Yidong]

As mentioned in an earlier email, we can now cut the CVS branch for
Emacs 23.  Any futher 23.0.xx pretest(s), and the 23.1 release, will
come from the branch.

As currently planned, we'll use the trunk for developing 23.2.  But if a
bugfix-only release proves necessary (e.g. if serious bugs come to light
after we release 23.1), we can change course and make 23.2 from the
branch while bumping the trunk release to 23.3, similar to the original
development plan for the Emacs 21 series.

Has anyone given thought to the branch tag?  If we follow previous
conventions, we'd name the branch EMACS_23_BASE, but this is somewhat
misleading.

Re: Branching

[Chong Yidong]

> Has anyone given thought to the branch tag?  If we follow previous
> conventions, we'd name the branch EMACS_23_BASE, but this is somewhat
> misleading.

I think what we want is to tag the current HEAD as EMACS_23_BASE, and
create the EMACS_23_1_RC branch from that tag.  Correct?

Re: Branching

[Glenn Morris]

Chong Yidong wrote:

>> Has anyone given thought to the branch tag?  If we follow previous
>> conventions, we'd name the branch EMACS_23_BASE, but this is somewhat
>> misleading.

That was a mistake, not a convention. :)

> I think what we want is to tag the current HEAD as EMACS_23_BASE, and
> create the EMACS_23_1_RC branch from that tag.  Correct?

EMACS_23_1_BASE and EMACS_23_1_RC would be consistent with 21.

 admin/make-tarball.txt says:

  cvs rtag EMACS_`NUMBER'_BASE
  cvs rtag -b EMACS_`NUMBER'_RC -r EMACS_`NUMBER'_BASE

If this is wrong, it should be corrected.

Re: Branching

[Miles Bader]

Glenn Morris <rgm@gnu.org> writes:
>>> Has anyone given thought to the branch tag?  If we follow previous
>>> conventions, we'd name the branch EMACS_23_BASE, but this is somewhat
>>> misleading.
>
> That was a mistake, not a convention. :)

Yeah, let's not repeat it... :O

-Miles

-- 
Selfish, adj. Devoid of consideration for the selfishness of others.

Re: Branching

[Chong Yidong]

Okay.  I made the EMACS_23_1_RC branch, coinciding with the
EMACS_23_1_BASE tag on HEAD.  The version number on the trunk is now
23.1.50.  (Hopefully, no mistakes in CVS usage this time round ;-)

To switch to tracking the branch in your CVS tree, do

  cvs up -r EMACS_23_1_RC

The branch will be used to release 23.1.  Until then, whenever you
commit a fix that is intended for inclusion in 23.1, please check it
into *both* the branch and the trunk.

(After the 23.1 release, it will no longer be necessary to commit to the
branch.  If, for whatever reason, we later decide to change course and
make a 23.2 bugfix-only release from the branch rather than the trunk,
we'll port the necessary fixes to the branch.)

The trunk is now open for all non-regression bugfixes.  New features can
also be added to the trunk, but please discuss on emacs-devel first.

Re: Branching

[Nick Roberts]

I think it's a good idea to put 

2009-06-21  Chong Yidong  <cyd@stupidchicken.com>

	* Branch for 23.1.

in the Changelogs so that people understand why some changes that have entries *before*:


	* Version 23.1 released.

are not included in the release (and can see what is in the release while on trunk).  I think
that I may have done this last time in src/ChangeLog.10 (in your name):


2007-04-24  Chong Yidong  <cyd@stupidchicken.com>

	* Branch for 22.1.

because I got confused about my GPM changes.

-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Branching

[Chong Yidong]

nickrob@snap.net.nz (Nick Roberts) writes:

> I think it's a good idea to put 
>
> 2009-06-21  Chong Yidong  <cyd@stupidchicken.com>
>
> 	* Branch for 23.1.
>
> in the Changelogs so that people understand why some changes that have
> entries *before*:
>
> 	* Version 23.1 released.
>
> are not included in the release (and can see what is in the release
> while on trunk).

Done, thanks.

Re: Branching

[Michael Albinus]

Chong Yidong <cyd@stupidchicken.com> writes:

> New features can also be added to the trunk, but please discuss on
> emacs-devel first.

I'm thinking about upgrade to Tramp 2.1.16. It contains a new package,
tramp-gvfs.el, which offers the new connection methods "obex", "dav" and
"davs". Internally, the GVFS fuse filesystem will be used for these new
methods, via D-Bus.

Any objections to install it in the trunk?

Best regards, Michael.

Re: Branching

[Chong Yidong]

Michael Albinus <michael.albinus@gmx.de> writes:

> I'm thinking about upgrade to Tramp 2.1.16. It contains a new package,
> tramp-gvfs.el, which offers the new connection methods "obex", "dav" and
> "davs". Internally, the GVFS fuse filesystem will be used for these new
> methods, via D-Bus.
>
> Any objections to install it in the trunk?

I have no objections.

Re: Branching

[Glenn Morris]

Chong Yidong wrote:

> Okay.  I made the EMACS_23_1_RC branch, coinciding with the
> EMACS_23_1_BASE tag on HEAD. 

I forget; do we need to make a Savannah support request to have
changes in this branch sent to the emacs-diffs mailing list?

Re: Branching

[Yavor Doganov]

Glenn Morris wrote:
> I forget; do we need to make a Savannah support request to have
> changes in this branch sent to the emacs-diffs mailing list?

Yes.  (But not this time, as I've just done that.)

Re: Branching

[Glenn Morris]

Yavor Doganov wrote:

> Yes.  (But not this time, as I've just done that.)

Thanks for the service!
(Procedure noted for future reference in admin/make-tarball.txt.)

Re: Branching

[Stefan Monnier]

> The branch will be used to release 23.1.  Until then, whenever you
> commit a fix that is intended for inclusion in 23.1, please check it
> into *both* the branch and the trunk.

Hopefully it's the last time we need this "commit to both" brain-damage.
Actually, maybe Miles still have his Arch gateway working so we could
automatically install the 23/1 patches into the trunk?


        Stefan