Plans for Python 3.0?

[protein_origami]

Hi,

I was wondering what's currently planned (or not planned) for making
boto work with python 3.0? Is that something folks are currently
working on? And if so, will that be for the 1.5 or 2.x branch of
boto?

I tried searching the group's message archives, but I didn't see
anything -- sorry if this is a FAQ.

On the off chance that a simple solution would work, I tried running
the "2to3" script on boto 1.5c, but it didn't seem to work quite
right.

Thanks,
Chris

[mitch]

Hi -

Originally, I was thinking the whole boto 2.0 effort could be focused
on Python 3.0. I'm not sure that's a good idea, though. I really
don't know how quickly people plan to adopt Python 3.0. It seems like
a significant effort to transition. Opinions?

In the meantime, I will fire up an EC2 instance and install Python 3.0
on it and see if I can get boto running with it.

Mitch

[mitch]

Just a quick followup. Based on about an hour spent installing Python
3.0, running the 2to3 utility and then looking at the results, my
initial conclusion is that getting boto to run in Python 3.0 will be a
non-trivial effort. Some of the things I found could probably be
handled automatically by improving the rules in 2to3 (e.g. doesn't
seem to handle mapping StringIO to io.StringIO, doesn't handle
ConfigParser -> configparser, etc.) but others, such as the changes
in moving httplib to http.client look like it will require
considerable manual effort.

I would like to get boto running in Python 3.0. The question is
when. Please respond to this thread if you have an imminent need to
run boto in Python 3.0.

Mitch

[protein_origami]

Hi Mitch,

I'm guessing I'm in the unfortunate minority on this one. I'm
starting a new
project, entirely from scratch, so I was hoping to go with python 3
from
the start and avoid the hassles of porting it later. Boto is the one
python module I really need. Oh well.

Thanks,
Chris

[mitch]

Well, I would like to get boto running in Python 3.0. I'm just not
sure exactly when I'll find the time to get it running and even then,
I'm sure it will take some time and testing to work out the kinks. If
you or others are willing to do that, that would be great.

I'm still trying to figure out the relationship between boto 2.0 and
Python 3.0. If I have spend time converting all of the httplib code
to http.client in Python 3.0, I'd kind of like to clean it up in the
process.

Mitch

[Eric Evans]

[ mitch ]

> Originally, I was thinking the whole boto 2.0 effort could be focused
> on Python 3.0. I'm not sure that's a good idea, though. I really
> don't know how quickly people plan to adopt Python 3.0. It seems like
> a significant effort to transition. Opinions?

If Guido's Pycon talk is any indication, then it will be a few years
before Python 3 is the gold standard.

You probably also want to wait until Python 3 is more widely distributed
by operating system vendors, (manual installations are pain, and in a lot
of cases, simply not possible), and until more third-party modules are
capable of running on 3.0.

Personally, I wouldn't be in any hurry. The 2.x series isn't going
anywhere for a long, long time.

--
Eric Evans
eev...@sym-link.com