Have you embraced Python 3.x yet?

[Harishankar]

Have you people embraced Python 3.x or still with 2.5 or 2.6?

I personally want to switch over but not too sure how many people are
using 3.x as opposed to 2 so I want to keep my programs compatible for
the majority.

-- Hari

 

[Alex Hall]

Because of compatibility, and many modules being built for 2.6 only, I
am still on 2.6.4 (updating to .5 soon).

 

[Neil Cerutti]

Have you people embraced Python 3.x or still with 2.5 or 2.6?

I personally want to switch over but not too sure how many
people are using 3.x as opposed to 2 so I want to keep my
programs compatible for the majority.

I switched 6 months ago, or so. My library needs are met by
what's in CPython, so I couldn't see, at the time, a good reason
to sit still.

My programs didn't take much effort to convert. They are nearly
all data conversion programs. Microsoft rendering the one outside
module I did use, an Excel reader, obsolete caused me a lot more
trouble than switching from 2.5 to 3.1.

Past advice in this forum has been that as long as you don't
depend on libraries that don't yet support Python 3, you can
probably switch over and not need to look back.

 

[Andrej Mitrovic]

I use both, really. I started by learning and using Python 3, but I
ended up using Python 2 much more often compared to Py3. Not because
of the functionality of the language, but because most software that
embeds Python or is extendable in some way is usually using Python 2.
+ There's the whole libraries issue (or the lack of them).

But I have to say, Python 3 is a sexy language, I can't wait for more
libraries to appear so I could focus more of my attention there.

 

[Kevin Walzer]

Have you people embraced Python 3.x or still with 2.5 or 2.6?

I personally want to switch over but not too sure how many people are
using 3.x as opposed to 2 so I want to keep my programs compatible for
the majority.

-- Hari

Not yet. Some key libraries I am using for apps under development
haven't been ported to support 3.x yet (feedparser, pycrypto).

 

[Gnarlodious]

I have been using Py3 since October, I switched over for the
OrderedDict feature. Some parts are a little hard to get used to, but
others are easier. I am doing web programming, so the UTF 8 default is
a big improvement.

-- Gnarlie

 

[Nobody]

Have you people embraced Python 3.x or still with 2.5 or 2.6?

Still with 2.6, and probably will be indefinitely.

I use Python mostly for Unix scripting: the kind of task which would
traditionally have used Bourne shell. For that purpose, life is much
simpler when everything is byte strings rather than character strings.

If I was doing the same thing on Windows, Python 3.x would probably make
more sense, as all of the OS APIs use Unicode (although there's still a
hell of a lot of software written using the "ANSI" interfaces; my
AppLocale folder has over a hundred entries).

 

[Mensanator]

Have you people embraced Python 3.x or still with 2.5 or 2.6?

3.1.

The only module I use regularly is gmpy and that's one that has
been updated.

 

[Terry Reedy]

Have you people embraced Python 3.x?

Yes. My current work only needs the stdlib and I much prefer the
streamlined language.

 

[geremy condra]

Have you people embraced Python 3.x or still with 2.5 or 2.6?

I personally want to switch over but not too sure how many people are
using 3.x as opposed to 2 so I want to keep my programs compatible for
the majority.

-- Hari

Yes- I write new code in 3.x and backport it PRN.

Geremy Condra

 

[David Cournapeau]

I won't switch until NumPy and SciPy make the jump.

We're almost there, though (mostly thanks to other people's work on Numpy):

http://github.com/cournape/scipy3/branches/py3k

David

 

[David Cournapeau]

Are you maintaining separate parallel Python 3 versus Python 2 code
bases?

No, we don't have the manpower to have two code bases - the C code
handles both versions, and the python 2 code is converted to python 3
"on the fly".

We are lucky not to have too much IO / string handling, which are the
hard things to handle with 2to3. The hard part was NumPy, because it
used the python C api so much. Porting scipy to something somewhat
usable was an half-day job,

cheers,

David