[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: python-django_1.8.18-1~bpo8+1_amd64.changes REJECTED




On May 25, 2017 8:09:23 AM EDT, Adrian Bunk <bunk@debian.org> wrote:
>On Thu, May 25, 2017 at 11:45:35AM +0000, Scott Kitterman wrote:
>> On May 25, 2017 6:53:49 AM EDT, Adrian Bunk <bunk@debian.org> wrote:
>> >On Thu, May 25, 2017 at 11:53:47AM +0200, Raphael Hertzog wrote:
>> >>...
>> >> The problem described here is that if we have Django reverse
>> >dependencies
>> >> in backports (I don't know if we have any), right now they have
>been
>> >> tested/validated with Django 1.7.x (stable) or 1.8.x
>> >(jessie-backports)
>> >> and switching to 1.10.x will likely break some of those packages
>that
>> >> were relying on deprecated features that got removed in 1.9.x and
>> >1.10.x.
>> >>...
>> >
>> >python-django 1.10 is supposed to have Breaks against all package
>> >versions that are or ever were in jessie or jessie-backports and
>> >that don't work with 1.10
>> >
>> >This would ensure that the user has working packages after
>> >  apt-get -t jessie-backports install python-django
>> >
>> >And it would also en passant give a list of rdeps that additionally 
>> >have to be backported when 1.10 gets backported.
>> 
>> 
>> That's irrelevant to the problem of 1.10 in backports.
>> 
>> The entire point of Django is Django applications.  It's impossible
>for non-trivial applications to go from Django 1.7 to 1.10 without
>modification.  This is almost all external to the archive and invisible
>to Debian.
>
>That's different from the problem of rdeps in Debian Raphael was
>describing.
>
>> If we care about this user code (and I think we should), then Django
>1.10 in backports is a non-starter.  If we don't care about user code
>then the python-django package isn't something I would ever recommend.
>
>The actual conclusion for what you are saying is that no Django feature
>
>release other than the one that will be shipped in the next stable 
>release of Debian should be uploaded to backports.
>
>In other words, Django 1.8 should not have been in backports at any
>time.
>
Sigh. No.

Scott K


Reply to: