Re: [Duplicity-talk] Python 3

[edgar . soldin]

On 21.08.2011 04:54, Michael Terry wrote:
> On 9 April 2011 16:19, Kenneth Loafman <address@hidden 
> <mailto:address@hidden>> wrote:
> 
>     The RedHat folks are on a 6-year support cycle, so the minimum Python we 
> need to support is 2.4 for now and will move up as time goes on.  It's a 
> pain, but we do have a fair number of them out in the wild, so we should not 
> drop support.  In my opinion, a 6-year old system is probably dangerous to 
> have around, even if patched, but some businesses do just that.
> 
> 
> In the course of testing some changes against Python 2.4, I was remembering 
> this conversation.  I looked up RHEL Python versions [1] and end-of-life 
> phases [2].
> 
> Which end-of-life phase do you want duplicity to target?
> 
> If it's the normal 7-year phase, that means we should support 2.4 through 
> early 2014 (and actually, that we should still be supporting 2.3).
> 
> There are other phases of end-of-life (4 and 5 years each) you could target 
> if you wanted to be more aggressive and a longer phase (10 years) if you 
> wanted to be more conservative.
> 
> [1] http://distrowatch.com/table.php?distribution=redhat
> [2] https://access.redhat.com/support/policy/updates/errata/
> 

my take is to keep these cycle in mind but to live the real life.

if there is a sound rationale for code requiring a more recent python we should 
consider it. if we can realize the same for older python versions without too 
much extra effort we should take the compatibility route.

e.g. http://docs.python.org/library/subprocess.html
used in backend.py . introduced in 2.4 it allows much cleaner process calls 
allowing to return stderr, stdout, exit code.

ee/duply.net