OK. So, I cloned a copy locally, ran 2to3, and started making fixes
for the remaining stuff that prevented it from even running. At this
point, 'light' mode seems to work OK, but there are occasional crashes
in the full mode, sometimes behavior depends on whether --chowconsole
or -d are used, as to how long it runs.

Several of the changes made after running 2to3 can probably just be
added to the trunk, regardless of Python version, and would help for
any future porting effort. Others may prove trickier and might require
the addition of some code to have slightly different behavior in
Python 2 vs Python 3. At any rate, one thing I wanted to ask was what
the minimum Python version Spyder supports (or will support in coming
releases), as that may have an impact on how some of the
incompatibilities would be dealt with.

I didn't do any work to add the automatic running of 2to3 in the build/
install process. I ran 2to3 on a branch, committed that, and then
worked from there to get it running. Hopefully most of the commits
after that can be added back into the mainline version, and enable
adding 2to3 to part of the build process, so as to allow Python3 work
without really hindering present development.

Should I make a clone in the google code space, and push commits there
for review/inclusion into mainline so that we can get closer to Python
3 support with 2to3?

Some of the changes I made were with regard to strings (legacy vs.
unicode), and they may not be entirely correct.  This is one aspect
that will probably take some more thought to enable Python 3 using
2to3 as part of the build process.

Oh, one other thing. I need to check if any crashing problems stem
from sys.std(err|in|out) usage when using pythonw on Windows. I found
some similar issues with IPython on Windows with Python 3. sys.std*
are None when using pythonw from Python 3, so anything that counts on
them being something else (builtins are fine, since they silently
ignore those streams when they're none).  Details at 
http://bugs.python.org/issue1415,
if interested.

Brandon

On Apr 16, 6:39 am, Pierre Raybaut <[email protected]> wrote:
> AFAIK, this is not a difficult job and there is no sticking point:
> it's just a question of time.
>
> So basically, this is 'only' a matter of:
> 1. running 2to3 on all sources
> 2. doing all the changes that the 2to3 tool can't do automatically
> 3. testing, testing, testing...
>
> -Pierre
>
> Le 16 avril 2012 12:04, Joes Staal <[email protected]> a écrit :
>
>
>
>
>
>
>
> > What would it take to support python 3? Numpy, scipy and  matplotlib all can
> > use python 3. Is it 'only' running 2to3 on all sources or are there more
> > intrinsic difficulties for getting spyder running with python 3?
>
> > Cheers,
>
> > Joes
>
> > --
> > You received this message because you are subscribed to the Google Groups
> > "spyder" group.
> > To view this discussion on the web visit
> >https://groups.google.com/d/msg/spyderlib/-/fqID4ngGUecJ.
>
> > To post to this group, send email to [email protected].
> > To unsubscribe from this group, send email to
> > [email protected].
> > For more options, visit this group at
> >http://groups.google.com/group/spyderlib?hl=en.

-- 
You received this message because you are subscribed to the Google Groups 
"spyder" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/spyderlib?hl=en.

Reply via email to