Python + Borland - msvcrt = grr.

[Stephen Hansen]

Hello,

    I'm wondering what compilers that you are aware of would throw such a
warning after something related to the following code. Borland is
complaining rather loudly (in the order of 200+ warnings, although some are
for other things I havn't looked into) because of the use of this idiom. I'm
not sugguesting it must be changed, but now that I understand what's going
on, it seems *correct* that this throws a warning.
    'retval' is assigned a value, in this case, NULL, and there is no path
of execution where it's even remotely possible for that NULL to be used...
so its a wasted operation. Why? Because retval is never actually used until
it is assigned to later on.

--- snip --- (from arraymodule, btw :))

static PyObject *
array_buffer_info(arrayobject *self, PyObject *args)
{
 PyObject* retval = NULL;
        if (!PyArg_ParseTuple(args, ":buffer_info"))
                return NULL;
 retval = PyTuple_New(2);
 if (!retval)
  return NULL;

 PyTuple_SET_ITEM(retval, 0, PyLong_FromVoidPtr(self->ob_item));
 PyTuple_SET_ITEM(retval, 1, PyInt_FromLong((long)(self->ob_size)));

 return retval;
}

--- snip ---

--Stephen
(replace 'nospam' with 'seraph' to respond in email)



"Fredrik Lundh" <fredrik at pythonware.com> wrote in message
news:nlXE6.9158$4N4.2042457 at newsc.telia.net...
> Stephen Hansen wrote:
> > It explains the infinate spam of warning messages I get, though,
> > since quite a few pointers are initialized to NULL and that NULL is
> > never read in. This is not something I think the Gods'o'Python
> > would care to have fixed, though. Dispite a 'no warnings' policy :)
>
> other compilers complain about possible "use before assignment" if
> you don't initialize them to NULL...
>
> Cheers /F
>
>