Patch for stricter implicit conversions between vectors

[Mark Shinwell]

Paolo Bonzini wrote:
>> - Disallows implicit conversions between generic vectors with differing
>>   element types, even if implicit conversions exist between those element
>>   types.
> 
> I would make this a bit more lax, allowing a conversion if the inner 
> element types have the same mode (e.g. vector char -> vector unsigned 
> char).

Conversions like that (especially signed <-> unsigned integers) were one of
the things this was deliberately designed to prevent -- see the original
thread :-)

Mark