duplicity-team team mailing list archive

[Kenneth Loafman <kenneth@xxxxxxxxxxx>]

On Sat, Mar 31, 2012 at 9:47 AM, Michael Terry <mike@xxxxxxxxxxx> wrote:

> I started a blueprint for this, to help break the work into smaller
> chunks: https://blueprints.launchpad.net/duplicity/+spec/python3
>
> -mt
>
> On 31 March 2012 09:28, Michael Terry <mike@xxxxxxxxxxx> wrote:
> > On 31 March 2012 07:40, Kenneth Loafman <kenneth@xxxxxxxxxxx> wrote:
> >> So, to get this straight, we may be able to supply a 3.2 patch in the
> Ubuntu
> >> release, yet supply 2.4 code for everyone else?  That is still a second
> >> version unless I'm missing something very subtle.  No real objections
> during
> >> the transition time, but I'll want to pull those together ASAP.
> >
> > OK sure, it would be a second version.  I'm just saying that I'd bear
> > the maintenance burden for it on the Ubuntu side until you were OK
> > with dropping RHEL 5.  At which point we can go back down to one
> > version.
> >
> > Oh and one thing that may be confusing.  When I said 2.6+ and 3.2+, I
> > meant with the same code base.  So we'd probably have some small
> > script that distributors/installers could use that changed the
> > /usr/bin/python at the top of any Python scripts to
> > /usr/bin/pythonX.X, for whatever version that distributor wanted to
> > use (RHEL would set to 2.6, Ubuntu would set to 3.2).  So we'd really
> > just be shipping 2.6+ code that happened to also work on 3.2+.  If
> > that's clear.
> >
> > And I wouldn't worry about ASAP.  There's no super urgency here unless
> > you want a good stretch of testing time.  But since the patch won't be
> > upstream for a bit, the testing will likely just happen for Ubuntu
> > users during the next development cycle.  Plus, Barry is making noise
> > about getting Python 2 off the CD for 12.10, but at UDS, the scope of
> > the work may cause him to push that goal back a cycle.  So actual
> > deadline for me here is still uncertain.
> >
> > -mt
>

Thanks for the good explanations and the blueprint.  This is going to be a
bit of work, but it will be worth it in the long run.

...Ken