[ros-users] REP 3: Target Platforms updates for Fuerte and Groovy

[Ken Conley]

On Mon, Nov 7, 2011 at 5:25 AM, Ingo Lütkebohle <iluetkeb at gmail.com> wrote:
> Hi,
>
> what about "official" Python 3 support, at least in the core libraries?

A general response to, "is X officially supported?"

'official' support usually means:

 1. There are tests/continuous integration for it
 2. Someone supports it

The TUM guys did a great job getting us a lot of the way there and
their patches have been incorporated, but #1 and #2 are still not
present.  The patches also don't cover command-line tools, just the
libraries exercised by MORSE.

Python 3 compatibility requires fairly complete coverage tests; I have
been refactoring the Python libraries to make this easier for someone
who wants to take this on, but there is no Py3k continuous
integration.  Py3k tests require a different set of assumptions
regarding strings, bytes, unicode, and iterators -- as an example, it
was possible to create filenames in 3.0 and 3.1 that Py3k programs
could not open.

Similarly, for #2, the general rules is that someone has to use the
thing being supported, as you can't support something you don't use.
We can provide resources, such as build farm resources, to someone who
does wish to take on this role.

We will continue to accept patches relating to Python 3, so long as
they don't break Python 2, and I will continue to update code to a
dual 2/3 style as I encounter it.  Trickier issues, such as the
bytes/string representation issues, remain active areas of work [1].

 - Ken

[1]: http://answers.ros.org/question/2032/smach-introspection-server-fails-in-electric?answer=4499#4499

> I'm asking specifically because we're currently using Blender and/or
> MORSE, which requires Python 3. The TUM guys have already done some
> work to port ROS messaging to Python 3, but as far as I know that is
> not official, yet.
>
> cheers,
> Ingo
>
> On Mon, Nov 7, 2011 at 10:17 AM, Ken Conley <kwc at willowgarage.com> wrote:
>> 2011/11/7 Stéphane Magnenat <stephane at magnenat.net>:
>>> Hello,
>>>
>>> What is the plan w.r.t. C++11? Now that the standard is out, and that a
>>> large part of features are available in recent g++ releases, we will see
>>> many upstream libraries going for it. As upstream developer, it is clearly
>>> always a question of whether to adopt such a recent standard, but given the
>>> huge speed-up in development it brings in some cases, I have decided to go
>>> for using features that are at least in gcc 4.4.
>>
>> The topic wasn't brought up during the Thirdparty SIG meeting.  The
>> approximation of that topic is that Lucid is the current lower
>> watermark (gcc 4.4) for integration in Fuerte and Groovy will move to
>> Oneiric (gcc 4.6).
>>
>> GCC C++ 11 status (still 'experimental'): http://gcc.gnu.org/projects/cxx0x.html
>>
>> It's always possible to reconvene the SIG to discuss C++ 11 more
>> specifically; I'm coordinator but I do not pretend to be knowledgeable
>> in that issue.
>>
>>  - Ken
>>
>>>
>>> To allow integration of recent code from research, I think that ROS should,
>>> whenever possible, embrace C++11.
>>>
>>> Kind regards,
>>>
>>> Stéphane
>>>
>>> --
>>> Dr Stéphane Magnenat
>>> http://stephane.magnenat.net
>>> _______________________________________________
>>> ros-users mailing list
>>> ros-users at code.ros.org
>>> https://code.ros.org/mailman/listinfo/ros-users
>>>
>> _______________________________________________
>> ros-users mailing list
>> ros-users at code.ros.org
>> https://code.ros.org/mailman/listinfo/ros-users
>>
>
>
>
> --
> Ingo Lütkebohle
> Bielefeld University
> http://www.techfak.uni-bielefeld.de/~iluetkeb/
>
> PGP Fingerprint 3187 4DEC 47E6 1B1E 6F4F  57D4 CD90 C164 34AD CE5B
> _______________________________________________
> ros-users mailing list
> ros-users at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>