Python 2.5.1 on x86_64 was: [Zope3-Users] Re: zopeproject install failure

[Tim Cook]

On Mon, 2007-09-17 at 14:01 +0200, Philipp von Weitershausen wrote:
> > There are numerous "warning: initialization from incompatible pointer
> > type" when compiling the .c modules.  
> 
> Yup. Mostly if not only from the ZODB.

There and from BTree/*


> > When I built Python 2.5.1 I had to use the CFLAG -fwrapv 
> > Is this the same issue?  If so, where can I add that flag so the zope .c
> > files will be compiled correctly?
> 
> What does that flag do? 

>From the GCC manual:

Using ‘-fwrapv’ means that signed overflow is
fully defined: it wraps. When ‘-fwrapv’ is used, there is no difference
between‘-fstrict-overflow’ and ‘-fno-strict-overflow’. With ‘-fwrapv’
certain types of overflow are permitted. For example, if the compiler
gets an overflow when doing arithmetic on constants, the overflowed value
can still be used with ‘-fwrapv’, but not otherwise.


> I can compile my stuff with Python 2.5.1 just fine.

Which GCC is it built with?  


> I've heard of another report where a MemoryError would occur with PYthon 
> 2.5 on Linux x86 64bit. I don't have access to such a machine, 
> unfortunately. It would be great if somebody with some C debugging 
> skills could check this out. It might already be good to know whether it 
> only occurs with zope.app.container, or also with other module that have 
> C extensions.

Probably any C code. 

> Note, this isn't a bug of zopeproject :). It just shows that Python 2.4 
> is still the only blessed platform, for a good reason.

You are correct and I shouldn't have used that subject line.  

I was ready to prove that it wasn't the Pythoon version but the GCC. So
I installed Python 2.4.4 using GCC 4.1.2 and expected the exact same
failures.  I was wrong.  It works as advertised.  Hmmmmmm.  

Now I'm really confused.  But; I haven't the skills to fix it.  This is
just so that others may see that even though some people are successful
with Zope 3.4.x / Python 2.5.1 it may be a platform dependency issue as
well.  

Cheers,
Tim



-- 
Timothy Cook, MSc
Health Informatics Research & Development Services
http://timothywayne.cook.googlepages.com/home
01-904-322-8582