This is from Evan Simpson, on the zope-dev list.  The 
bulk have to do with what happens if __name__ is set to 
None:


Subject: Re: [Zope-dev] Re: Zope 2.7.0 rc2 + python 
2.3.3 problem
http://mail.zope.org/mailman/listinfo/zope-dev

Python 2.3.3 (#2, Jan 13 2004, 00:47:05)
[GCC 3.3.3 20040110 (prerelease) (Debian)] on linux2
Type "help", "copyright", "credits" or "license" for more 
information.
 >>> range(1.0)
__main__:1: DeprecationWarning: integer argument 
expected, got float
[0]
 >>>

So far, so good.  However:

Python 2.3.3 (#2, Jan 13 2004, 00:47:05)
[GCC 3.3.3 20040110 (prerelease) (Debian)] on linux2
Type "help", "copyright", "credits" or "license" for more 
information.
 >>> __name__=None
 >>> range(1.0)
[]
 >>> 1+1
Traceback (most recent call last):
   File "/usr/lib/python2.3/warnings.py", line 57, in warn
     warn_explicit(message, category, filename, lineno, 
module, registry)
   File "/usr/lib/python2.3/warnings.py", line 63, in 
warn_explicit
     if module[-3:].lower() == ".py":
TypeError: unsubscriptable object
 >>>

...and...

Python 2.3.3 (#2, Jan 13 2004, 00:47:05)
[GCC 3.3.3 20040110 (prerelease) (Debian)] on linux2
Type "help", "copyright", "credits" or "license" for more 
information.
 >>> import warnings
 >>> warnings.simplefilter("error", 
category=DeprecationWarning)
 >>> range(1.0)
[]
 >>> 1+1
Traceback (most recent call last):
   File "/usr/lib/python2.3/warnings.py", line 57, in warn
     warn_explicit(message, category, filename, lineno, 
module, registry)
   File "/usr/lib/python2.3/warnings.py", line 92, in 
warn_explicit
     raise message
DeprecationWarning: integer argument expected, got 
float
 >>>

...and...

Python 2.3.3 (#2, Jan 13 2004, 00:47:05)
[GCC 3.3.3 20040110 (prerelease) (Debian)] on linux2
Type "help", "copyright", "credits" or "license" for more 
information.
 >>> import pdb
 >>> __name__ = None
 >>> pdb.run('range(1.0)')
 > <string>(1)?()
(Pdb) s
--Call--
 > /usr/lib/python2.3/warnings.py(24)warn()
-> def warn(message, category=None, stacklevel=1):
(Pdb) r
--Return--
/usr/lib/python2.3/bdb.py:302: RuntimeWarning: 
tp_compare didn't return 
-1 or -2 for exception
   i = max(0, len(stack) - 1)
[traceback snipped]

Looks like something isn't properly propagating 
exceptions.

Cheers,

Evan @ 4-am

Logged In: YES 
user_id=6656

Might this be specific to range()?

[mwh@pc150 build]$ ./python.exe -Werror
Python 2.4a0 (#3, Feb  3 2004, 19:23:25) 
[GCC 3.3 20030304 (Apple Computer, Inc. build 1493)] on darwin
Type "help", "copyright", "credits" or "license" for more 
information.
>>> range(5.0)
[]

range()'s argument handling is somewhat odd

Logged In: YES 
user_id=6656

OK, the problem is that returning NULL from 
getargs.c:convertsimple indicates *success* (argh!).

Fixing that provokes weird errors from handle_range_longs.

Logged In: YES 
user_id=6656

How's this?  It's horrible, but I'm not sure that can be avoided.

Logged In: YES 
user_id=6656

Oops, wrong file.

Logged In: YES 
user_id=6656

Is anyone listening?

Logged In: YES 
user_id=31435

I'm listening, but with half of part of one ear.  Have to agree 
convertsimple() was wrong in these cases, but can't make 
time for more than that.  Reassigned to Jeremy, partly at 
random (his is one of the names that shows up as a recent 
getargs.c changer).

Logged In: YES 
user_id=1188172

In current SVN heads, range(1.0) gives DeprecationWarning and 
__name__=None; range(1.0) gives TypeError.

Is this okay?

Logged In: YES 
user_id=31435

No, it's certainly not OK for range(1.0) to raise either
DeprecationWarning or TypeError depending on what __name__
happens to be.  But I may not understand what you mean. 
This is a screen scrape from current SVN trunk, on Windows:

>>> __name__ = None; range(1.0)
None:1: DeprecationWarning: integer argument expected, got float
[0]

I don't see TypeError there, so I'm not sure what

"""
In current SVN heads, range(1.0) gives DeprecationWarning and 
__name__=None; range(1.0) gives TypeError.
"""

means.

Logged In: YES 
user_id=1188172

I fixed warnings inbetween, see revision 42028.

So I guess this can be closed.