[IPython-dev] Testing needed: python 2.3 installation

[Fernando Perez]

On Sat, May 3, 2008 at 9:18 AM, Ville M. Vainio <vivainio at gmail.com> wrote:
> I made a change in setup.py (in bzr trunk) that uses package_data
>  (supported in setuptools, and distutils versions shipped with python
>  2.4+) to ship UserConfig folder, in order to unbreak easy_install.
>
>  Could someone that has python 2.3 installed try doing "python setup.py
>  install", delete/move away ~/.ipython and see that ipython starts up
>  normally?

Well, it looks like ipython hasn't worked with 2.3 for a while:

bic128[bin]$ ./ipython
Traceback (most recent call last):
  File "./ipython", line 27, in ?
    IPython.Shell.start().mainloop()
  File "/home/fperez/tmp/local//lib/python2.3/site-packages/IPython/Shell.py",
line 1215, in start
    shell = _select_shell(sys.argv)
  File "/home/fperez/tmp/local//lib/python2.3/site-packages/IPython/Shell.py",
line 1157, in _select_shell
    all_opts = set(['tk','pylab','gthread','qthread','q4thread','wthread',
NameError: global name 'set' is not defined

The real question is: have we reached the point where we should just
declare 2.4 to be our base requirement?

I'm happy to continue supporting 2.3 (the above problem is trivial to
fix, though there may be more), but 2.4 has been out for 3 1/2 years
already.  We have painfully limited developer time and plenty of
not-done work.  Unless someone makes a good case for why 2.3 support
is key, I'm inclined to dump it for the next release of ipython0.
IPython1 already depends on it anyway (and some of the new goodies
we're putting in actually need 2.5!).

Cheers,

f