web2py and Python3

[Álvaro Justen [Turicas]]

Hello all,
is there a timeline to web2py be supported in Python3?
I think this information important for who is studing and choosing a
web framework, like me.
Django, for example, will support Python3 in 2010.

--
Álvaro Justen
Peta5 - Telecomunicações e Software Livre
21 3021-6001 / 9898-0141
http://www.peta5.com.br/

[mdipierro]

This issue is not "porting". The issues are:

1) the database drivers have to be ported first. web2py and Django
depend more or less on the same drivers

2) web2py supports the Google App Engine and that is Python 2.5. If we
move to 3.0 your applications will not run anymore on the GAE. Google
has no plan to move to 3.0.

3) A framework is not just based on Python. It requires that you code
in Python. If people "port" Django to Python 3.0. Your Django
applications will no longer work. The same is true for web2py.

The bottom line is that we could port web2py to 3.0 in one week
(technically it does not take much, except for str vs unicode we are
already compliant) but it would be pointless to lose driver support,
GAE support and break compatibility of all the current apps.

When GAE moves to 3.0 and the database drivers for all supported
backends become available we will release something like web3py (TM).
Since we are going to break language backward compatibility that will
also be a good time to include other non-backward compatible changes.
2010-2011 are reasonable dates but just a guess.

You should not measure progress of a framework from this deadline, but
from the number of commits.

What feature of 3.0 do you consider so important to constraint you?

Massimo



On Apr 4, 8:58 pm, Álvaro Justen [Turicas] <alvarojus...@gmail.com>
wrote:

[Michal Jursa]

>
> What feature of 3.0 do you consider so important to constraint you?
>
> Massimo
>

LOL, I wanted to ask the same question..:))
+1

Michal alias Plysak

[Álvaro Justen [Turicas]]

On Sun, Apr 5, 2009 at 12:07 AM, mdipierro <mdip...@cs.depaul.edu> wrote:
>
> This issue is not "porting". The issues are:
>
> 1) the database drivers have to be ported first. web2py and Django
> depend more or less on the same drivers
>
> 2) web2py supports the Google App Engine and that is Python 2.5. If we
> move to 3.0 your applications will not run anymore on the GAE. Google
> has no plan to move to 3.0.
>
> 3) A framework is not just based on Python. It requires that you code
> in Python. If people "port" Django to Python 3.0. Your Django
> applications will no longer work. The same is true for web2py.
>
> The bottom line is that we could port web2py to 3.0 in one week
> (technically it does not take much, except for str vs unicode we are
> already compliant) but it would be pointless to lose driver support,
> GAE support and break compatibility of all the current apps.
>
> When GAE moves to 3.0 and the database drivers for all supported
> backends become available we will release something like web3py (TM).
> Since we are going to break language backward compatibility that will
> also be a good time to include other non-backward compatible changes.
> 2010-2011 are reasonable dates but just a guess.
>
> You should not measure progress of a framework from this deadline, but
> from the number of commits.
>
> What feature of 3.0 do you consider so important to constraint you?

Specifically, none. I only would want to know about the future of
web2py to know when I'll have to "invest" in porting to Python 3.

[AchipA]

It's a vicious circle of apps (frameworks) and libs not upgrading
because of the other.

What we *could* do is to make everything web3py friendly *without*
separating the 2 and 3. Have a 2.6 base ready that can be put through
2to3 (and perhaps a custom web2to3py script) that transforms web2py
into web3py to allow people to write python3 apps even though the
majority is still on 2.x and not all database drivers are there yet.
Why bother you say when (apart from unicode and longints) you gain no
significant features ? Because this can give us an edge over other
frameworks both in terms of tech and in terms of PR. The more time
passes, the more people will use any particular framework and make it
all the harder to migrate to python3. As the Django guy on the ORM
panel said - we'll be the last ones there.

web2py with it's batteries included style self-containedness is a
prime candidate to make that jump first.

[mdipierro]

I think the only issue is str vs unicode and it depenends on how the
handle it in the drivers. everything else should be 3 friendly aready.