Thanks for the reference and info, Daniele. Yes, clang on my end. I've been able to connect to my db and
interact with my plpgsql functions okay, so no trouble yet. But I will research a bit further.
Now I just have the minor detail of re-absorbing Python as I haven't written any since about 2006!
It's all coming back…
Again, big thanks as I am sure I would still be banging my head if you hadn't pointed out that I was
using the wrong install program/version since that would have been too simple and obvious a solution
to consider. Ha!
- Leon
On Apr 1, 2012, at 9:46 AM, Daniele Varrazzo wrote:
> On Sun, Apr 1, 2012 at 1:54 AM, Leon Starr <leon_starr@modelint.com> wrote:
>> Thanks, Daniele, you've saved my weekend! (I think)
>
> That's great news :)
>
>
>> BUT, there were a few dozen unsettling warnings generated during the
>> easy_install and I am wondering if this is just
>> a 2->3 compatibility thing. How concerned should I be about this kind of
>> stuff?
>
> It's a false positive, don't know what compiler is on your system,
> probaby clang. I've tried enabling -Warray-bounds on gcc but I don't
> get such warnings. I've found discussions about clang -Warray-bounds
> being over-zealous:
> <http://comments.gmane.org/gmane.comp.compilers.clang.devel/15813>.
>
> The Python tuple structure is defined with a 1-item array ob_item at
> the end, but it is initialized with extra trailing memory so that the
> other array items "hang off" the structure. It's a classic way of
> defining flexible arrays in C structures. And anyway it's in Python
> code, not in Psycopg's.
>
> Cheers,
>
>
> -- Daniele
