[ckan-dev] release branches, etc

Mon Feb 28 17:59:58 UTC 2011

On 28 February 2011 17:43, Seb Bacon <seb.bacon at gmail.com> wrote:
> Hi,
>
> I'm working on a client deployment (DataGM).  The currently
> checked-out branch on the live server is ckan-1.3; it used to be
> metastable.  Someone else set it up & I'm not sure what the
> best-practice here is.  In particular, there are some bugs recently
> fixed in various branches and I'm not entirely clear of the process
> whereby they get into release branches.
>
> Do we have a documented policy somewhere on this?

This a roughly accurate state of play, although there is still plenty
of discussion about details as things happen:
http://trac.ckan.org/wiki/BranchingPolicy

1.3 is definitely a 'supported release' (1.2 is perhaps questionable)
so that any bugs we should fix on the 1.3 branch. So do use 'hg
transplant' (or simply 'hg export {rev} >file' and 'hg patch file' to
put the fix in. Then it's normal to 1.3 back into default (expect a
nop in this case, but that is useful to know).

BTW I've set the version number on that branch to be '1.3.1' to show
on a live instance when it is the same as the 1.3 released to pypi, or
has fixes. I also suggested to Rufus and David Raznick on Friday to
call an interim release 1.3.1, so maybe that should be 1.3.2 or even
1.3.5 since we're almost halfway to the scheduled release of 1.4. But
although this particular release branch will probably have the fix you
wanted, it also has lots of new features, and has not yet had the
week's stabilization period on ckan.net that 1.3 has had.

It's always tempting to say 'just use the latest bleeding edge
release', but we should encourage each other to stick to our process
designed to reduce risks, particularly now we've been forced out of
beta in Manchester by the Guardian...

Dave

>  If I want to
> cherry-pick bugfixes to the ckan-1.3 release branch, what should I do
> / who should I ask / etc?

>
> As an example, I want this fix:
> https://bitbucket.org/okfn/ckan/changeset/1b8fedeb7ab0/
>
> Thanks
>
> Seb
>
> _______________________________________________
> ckan-dev mailing list
> ckan-dev at lists.okfn.org
> http://lists.okfn.org/mailman/listinfo/ckan-dev
>