[PKG-Openstack-devel] Bug#812309: Bug#812309: Bug#812309: python-trollius: upgrade to version 2.0

[Thomas Goirand]

On 02/09/2016 09:42 PM, Sandro Tosi wrote:
> On Fri, Feb 5, 2016 at 2:52 PM, Thomas Goirand <zigo at debian.org> wrote:
>> On 02/05/2016 05:23 PM, Sandro Tosi wrote:
>>>> I've been working on this, and there's an issue when running unit tests
>>>
>>> i see you just pushed the changes for 2.0 (they were not there when I
>>> sent my emails/bug)
>>
>> Yes, and the version 2.1~b1 currently in the Git should work. Though I
>> still prefer to wait for upstream to release before uploading to Debian.
>> Hopefully, this will happen today. If not, I'll upload 2.1~b1.
> 
> any idea why the release has been delayed? just lack of time or some
> tech issues?

As much as I understand, what's taking so long is that upstream insists
on providing Windows wheels, for win32 and win64, and with python 2.7,
3.3 and 3.4. This means 6 builds, which all have to be done with a
different build environment.

I've been looking and waiting for this version 2.1 to be released, but
it's still not coming. As you're waiting for it, and I don't want to
make you wait forever (especially considering it will take 5 days to
migrate to Testing before you'll be able to upload to backports), I
decided not to wait, and I uploaded 2.1~b1-1 to unstable, as normally,
the final release wont be any different.

> Also, I would like to have that version in Jessie: would you like to
> take care of the backport yourself or prefer me to do it?

Feel free to maintain the backport if you like. For the packages I
maintain, the older 1.0.1 which is already in Stable is enough, so I
have no interest in maintaining a backport for a newer version. I
maintain enough packages in Debian, so it's ok to give out some.

BTW, could you let me know why you need Trollius in Jessie-backports?
The fact that upstream declared it as deprecated, and will not do any
work on it but basic maintenance (which as you saw, he's not in a hurry
to do) isn't very engaging. I would therefore recommend to be very
cautious with this package.

I hope other dependencies in OpenStack will get rid of it, but probably
we need to wait that everything switches to Python3 (in which case they
can use asyncio directly), which is going to take probably a very long
time, but that's still the plan. Then I may either orphan or ask for
this package to be RM, unless a new person becomes the upstream. I'll
care to ping you when/if this happens. You've been warned! :)

Cheers,

Thomas Goirand (zigo)