[ZODB-Dev] ExtensionClass tp_compare fix

[Brian Lloyd]

> Thanks, Tim - good to know that I haven't lost my senses completely over
> the year end -:)...
> 
> Yes, that's the warning I'm seeing, and '$Id: ExtensionClass.c,v 1.58
> 2002/11/12 20:04:20 jeremy Exp $'
> is the version I got as well.
> 
> Hope Brian can check on the installer, as I'm not experienced in
> building from sources.
> 
> Cheers, Martin

Sorry, I'm way behind on mail. There is definitely a problem 
with the new (for 2.7) installer that I need to work out. It 
seems to have built beta 2 but labelled it b3 :(



Brian Lloyd        brian at zope.com
V.P. Engineering   540.361.1716              
Zope Corporation   http://www.zope.com 

> 
> -----Original Message-----
> From: Tim Peters [mailto:tim at zope.com] 
> Sent: Monday, 05 Jan 2004 19:31
> To: Gfeller Martin; Brian Lloyd
> Cc: zodb-dev at zope.org
> Subject: RE: [ZODB-Dev] ExtensionClass tp_compare fix
> 
> 
> [Martin Gfeller, trying Zope 2.7b3 on Windows]
> > Thanks, Tim.
> >
> > Now I'm a bit confused - I thought I had downloaded and installed B3,
> > but its Control Panel says it's B2 - although the Python build it
> > reports is 2.3.2 (#49, Oct 3 2003, 10:25:57), which is after the
> > release date of B2 (I didn't downloade Python separately, and am
> > using the binary build).
> >
> > ExtensionClass.pyd has the date of 03 Oct 2003, 10:21:41.
> >
> > Could it be that B3 doesn't know how advanced it is :-) ?
> > But that would mean that I still have the problem.
> 
> Thanks for the followup, Martin.  I don't know how the Zope 2.7 beta 3
> installer got built for Windows, but I agree it looks goofy in at least
> the
> ways you've reported.  I downloaded it just now and installed it:
> 
>     Zope-2.7.0-b3-win32.exe
> 
> Control Panel indeed says
> 
>     Zope Version  (Zope 2.7.0-b2, python 2.3.2, win32)
> 
> so maybe the 2.7b3 Windows installer in fact packaged 2.7b2 again (or
> even
> something older than that -- see below).
> 
> >From the 2.7B3 *tarball* download, I verified that the fix for
> ExtensionClass comparisons that started this is in what Linux people
> think
> of as 2.7b3.  But I don't think it's in the ExtenionClass binary
> installed
> by the 2.7b3 Windows installer:
> 
> C:\Program Files\Zope-2.7.0-b3>set PYTHONPATH=lib/python
> 
> C:\Program Files\Zope-2.7.0-b3>bin\python
> Python 2.3.2 (#49, Oct  3 2003, 10:25:57) [MSC v.1200 32 bit (Intel)] on
> win32
> Type "help", "copyright", "credits" or "license" for more information.
> >>> import ExtensionClass
> >>> print ExtensionClass.__doc__
> ExtensionClass - Classes implemented in c
> 
> Built-in C classes are like Built-in types except that
> they provide some of the behavior of Python classes:
> 
>   - They provide access to unbound methods,
>   - They can be called to create instances.
> 
> $Id: ExtensionClass.c,v 1.58 2002/11/12 20:04:20 jeremy Exp $
> 
> >>>
> 
> The last line of text there says the ExtensionClass version actually
> installed is CVS revison 1.58, more than a year old.  That would explain
> why
> you're seeing problems that were fixed a long time ago (but not *that*
> long
> ago <wink>).  Continuing the session above:
> 
> >>> class A(ExtensionClass.ExtensionClass):
> ...   def __init__(self):
> ...     pass
> ...
> >>> a = A()
> >>> b = A()
> >>> cmp(a, b)
> __main__:1: RuntimeWarning: tp_compare didn't return -1, 0 or 1
> -1
> >>>
> 
> That's the specific RuntimeWarning you're seeing, right?
> 
> 
> I'm copying Brian Lloyd in the hope that he can shed more light on this.
> 
>