[Debian-science-sagemath] Removal of python-rpy2

Ximin Luo infinity0 at debian.org
Tue Sep 5 08:05:00 UTC 2017


Laurent Gautier:
> 2017-09-03 10:54 GMT-04:00 Tobias Hansen <thansen at debian.org>:
> 
>> On 09/03/2017 02:28 PM, Dirk Eddelbuettel wrote:
>>> | Is it now suddenly impossible to provide python-rpy2?
>>>
>>> I am afraid so.
>>>
>>> | Could you please bring it back?
>>>
>>> That would be a departure from upstream I do not think serves anyone.
>>>
>>> Can you within sagemath add it as a local component?
>>
>> I think the way to go is to introduce a new package rpy2-2.8 that builds
>> only python-rpy2 for Python 2. That's not deviating from upstream and
>> sagemath is using rpy2 2.8 anyway. Do you want to maintain this or shall
>> we (the sagemath team)?
>>
> 
> Just a note: I don't know how well Python will handle several versions of
> the same package.
> 

Python 2 and 3 are completely separate so it should be fine to maintain and allow co-installation of rpy2 2.8 for Python 2 but the latest version for Python 3.

>> On 09/03/2017 02:34 PM, Jerome BENOIT wrote:
>>> it sounds as a race issue.
>>>
>>> When SageMath will switch to Python 3 ?
>>
>> They are working on it and they know that other projects have announced
>> to stop supporting Python 2, for example on http://www.python3statement.
>> org/
>> That page also has a chart that lists rpy2 2.8 as supported during 2018
>> with Python 2 support.
>>
> 
> There is no plan to remove older rpy2-2.8.x releases on pypi. The support
> for Python 2 means that critical bugs will be fixed and minor releases in
> the 2.8.x series will be made. 2018 is coming soon though, and I'll likely
> accept pull requests after that date.
> 
> 

Then this implies that we should maintain 2.8 in Debian in case anyone needs to use it in Python 2, which is the case for SageMath. Separately, we could maybe ask them to stop depending on it.

X

-- 
GPG: ed25519/56034877E1F87C35
GPG: rsa4096/1318EFAC5FBBDBCE
https://github.com/infinity0/pubkeys.git



More information about the Debian-science-sagemath mailing list