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

Re: Python 3.4 backport



Thanks for the feedback Dimitri. I now understand the motivation for the lack of Python backports. It was as I suspected, being such an infrastructural part of the system. Going with a private repository and compiling the unstable source package in Wheezy will meet my needs.

--Pedro.


On 17 May 2014 00:24, Dimitri John Ledkov <xnox@debian.org> wrote:
On 16 May 2014 16:15, Pedro Romano <pedro@salesseek.co.uk> wrote:
> I haven't been able locate a backport of Python 3.4 to Wheezy, and so I am
> considering tackling the backport myself. I would have expected that Python
> 3.4 to be an obvious candidate for a backport, so is there any reason
> preventing the backport?
>

Major python releases are not obvious candidates for backports.

1) all compiled extensions need to be recompiled to pick up python3.4,
otherwise only standard-library will be available
2) packages had to be ported to support 3.4 (e.g. changes such as
not-inheriting file descriptors did cause a few subtle, but critical
runtime errors), thus mixing python3-* modules from wheeze with
python3.4 will result in errors
3) not everything in sid has been ported to python3.4 yet, as it's not
yet the default version

What do you require from 3.4 that cannot be achieved with 3.2?
Why would you want to run completely unsupported combinations of software stack?

If you really require 3.4, I'd recommend running/using it from an
unstable/testing chroot created on stable host. Or otherwise use any
other container-like solution (virtual-envs, lxc containers, etc...)

If you can, just upgrade to testing and enjoy python3.4.

--
Regards,

Dimitri.



--
Pedro Romano
Software Engineer
SalesSeek Limited

Registered in England No. 08235398
Registered office: Unit 10, Turnham Green Terrace Mews, London, W4 1QU

Reply to: