Re: backport python-astropy for jessie?

["Leo P. Singer" <leo.singer@ligo.org>]

> On Jan 26, 2017, at 4:21 AM, Andreas Tille <tille@debian.org> wrote:
> 
> On Thu, Jan 26, 2017 at 10:00:43AM +0100, Ole Streicher wrote:
>>> I admit I'm not very motivated to sponsor backports.  I somehow need to
>>> draw a line what I can do and what not.  However, I'd recommend to read
>>> the backport guidelines.  As far as I remember its not necessary to be
>>> a DD but rather to get accepted by the backport team.
>> 
>> We don't have a specific policy for Debian Astro and mainly use the
>> Debian Science policy as a guide (and I insist in using git).
> 
> May be it makes sense to copy the Backporting paragraph to Debian
> Science.
> 
>> So, it
>> would probably wise to follow Debian-Meds guide to put the backports
>> into a "jessie-backports" branch of the reporitories, and to follow
>> https://backports.debian.org/Contribute/
> 
> The branch is actually called
> 
>    debian/jessie-backports
> 
> we used only jessie-backports but somebody raised some good arguments
> (which I forgot) to prefix it with debian/.

Hi Andreas,

Git will not let me name the branch 'debian/jessie-backports' because there is already a branch called 'debian':

$ git branch -m debian/jessie-backports
error: 'refs/heads/debian' exists; cannot create 'refs/heads/debian/jessie-backports'
fatal: Branch rename failed

Apparently, it is a known issue with git that you cannot create a branch 'foo/bar' if branch 'foo' exists:

http://stackoverflow.com/questions/22630404/git-push-refs-heads-my-subbranch-exists-cannot-create

Since it looks like most debian-astro packages have a 'debian' branch rather than 'master', is it OK if for debian-astro we instead adopt the practice that backports go in the 'jessie-backports' branch?

Thanks,
Leo