IntColumn raises TypeError on assignment of an mpz

So, the problem code is in IntVariable:

   def parse_set(self, value, from_db):
        if not isinstance(value, (int, long, float, Decimal)):
            raise TypeError("Expected int, found %r: %r"
                            % (type(value), value))
        return int(value)

The intent of this code is to only accept "number" type values, rather than anything that has an __int__ that succeeds. For instance, we want the following to raise an error:

    foo.int_column = '42'

I'm not sure of what the best course of action is. I don't think we want to be checking for every possible number-like type that has ever been created individually.

On Wed, 26 Jan 2011 21:45:49 -0000, James Henstridge <email address hidden> wrote:
> So, the problem code is in IntVariable:
>
> def parse_set(self, value, from_db):
> if not isinstance(value, (int, long, float, Decimal)):
> raise TypeError("Expected int, found %r: %r"
> % (type(value), value))
> return int(value)
>
> The intent of this code is to only accept "number" type values, rather
> than anything that has an __int__ that succeeds.

Strings don't have __int__s:

>>> 'a'.__int__
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'str' object has no attribute '__int__'

Rather the conversion is done by code in int.__new__ (effectively, at
least).

> For instance, we want the following to raise an error:
>
> foo.int_column = '42'
>
> I'm not sure of what the best course of action is. I don't think we
> want to be checking for every possible number-like type that has ever
> been created individually.

I think checking for __int__ is OK, something like:

try:
    int_conv = value.__int__
except AttributeError:
    raise TypeError(...)
else:
    return int_conv()

Checking for a number type is one of those things that's frustratingly
much easier in C than in Python!

Cheers,
mwh

I had forgotten the detail over where the string->integer conversion occurred, but I was pretty sure that was the reason for the check.

Checking for __int__ looks like a good replacement. Does it handle the case you are interested in?